System

Abstract

A system is provided.SOLUTION: A system, comprising: means for receiving an ambiguous condition input by a user and sending it to a server; means for the server to parse the received ambiguous condition based on a generative AI model and convert it into an appropriate query; means for the server to crawl a plurality of e-commerce platforms and aggregate product information based on the converted query; and means for the server to filter and rank the aggregated product information based on the user's specified condition.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] In today's e-commerce environment, many users often search for products based on vague requirements, without specifying specific product names or detailed conditions. However, conventional systems have difficulty providing search results that accurately reflect users' vague requirements, and manually searching multiple e-commerce platforms is time-consuming. Furthermore, only a limited number of systems have the functionality to analyze users' past behavior and preferences and make personalized product suggestions. A system that can solve these issues is needed. [Means for solving the problem]

[0005] The present invention provides a system that receives vague conditions entered by a user and transmits them to a server, analyzes the received vague conditions based on a generative AI model and converts them into an appropriate search query, crawls multiple e-commerce platforms and aggregates product data based on the converted search query, filters the aggregated product data and ranks it based on the user's specified conditions, and transmits the filtered and ranked product data to a terminal, which displays it to the user. Furthermore, the system includes a means for the server to analyze the user's past purchasing behavior and search history and make personalized product suggestions, and a means for the terminal to transmit user feedback to the server and form a feedback loop to improve the accuracy of the next search, thereby realizing a system that quickly and accurately provides products desired by the user.

[0006] "User" means an individual or organization that uses the System to search for products.

[0007] "Device" refers to the device (smartphone, tablet, PC, etc.) that a user uses to enter search criteria and display search results.

[0008] A "server" is a central computing system that receives input from users, analyzes conditions, collects data, and filters and ranks results.

[0009] "Vague conditions" are abstract requests or wishes that users enter without providing specific product names or detailed information.

[0010] A "generative AI model" is a model that uses artificial intelligence (AI) technology to analyze ambiguous input conditions and convert them into appropriate search queries.

[0011] An "e-commerce platform" is a website or service (such as Amazon, eBay, or Rakuten) that allows people to buy and sell goods and services over the internet.

[0012] "Crawling" is the process of automatically gathering information from a website or database.

[0013] A "search query" is a specific search term or phrase that a generative AI model generates by analyzing a user's vague criteria.

[0014] "Product Data" means product information (such as price, description, and rating) collected from the e-commerce platform.

[0015] "Filtering" is the process of selecting aggregated product data that meets specific criteria.

[0016] "Ranking" is the process of ordering filtered products based on criteria such as price, popularity, and review rating.

[0017] A "feedback loop" is a mechanism for continuously collecting feedback from users to improve the system's search accuracy and the quality of personalized suggestions. [Brief explanation of the drawings]

[0018] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6]FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0019] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0020] First, the terms used in the following description will be explained.

[0021] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0022] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0023] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0024] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0025] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0026] [First embodiment]

[0027] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0028] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0029] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0030] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0031] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0032] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0033] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0034] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0035] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0036] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0037] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0038] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0039] In the embodiment of the present invention, a system is provided that allows users to easily search for products they want and obtain optimal results. A specific embodiment of this system and its operation will be described below.

[0040] Getting User Input

[0041] User

[0042] Users enter vague criteria into the search bar on their devices (smartphones, tablets, PCs, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[0043] Terminal

[0044] The terminal accepts information entered by the user, converts the data into an appropriate data format such as JSON, and sends it to the server.

[0045] Condition Analysis

[0046] server

[0047] The server receives vague conditions sent from the device. This data is passed to the generative AI model. The generative AI model analyzes the user's vague conditions and converts them into a specific search query. This converted search query is optimized to search for the product the user is looking for.

[0048] Crawling e-commerce sites

[0049] server

[0050] The server crawls multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.) using the search query generated by the generative AI model. It sends the search query to each platform and retrieves relevant product data, including prices, descriptions, user reviews, and ratings.

[0051] Filtering and ranking results

[0052] server

[0053] The server aggregates product data obtained from each e-commerce platform. The aggregated product data is filtered and narrowed down based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[0054] Providing optimal search results

[0055] server

[0056] The server transmits the filtered and ranked product data to the terminal.

[0057] Terminal

[0058] The device displays the received product data to the user in a list view, grid view, or other display format, allowing the user to easily find the best product from a wide range of options.

[0059] Personalized recommendations and feedback loops

[0060] server

[0061] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to improve the accuracy of the next search criteria. This provides the user with more personalized search results and product suggestions.

[0062] user

[0063] Users can select a product from the product list displayed on their device and proceed to the details page or purchase page. They can also enter new search criteria and search again.

[0064] Specific examples

[0065] For example, if a user who wants to find autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete," the system will perform the following actions.

[0066] 1. User Input

[0067] The user types "I want autographed merchandise from a famous athlete" into the device.

[0068] 2. Condition analysis

[0069] The server passes the input to a generative AI model for analysis, resulting in the search query "famous athlete signed merchandise."

[0070] 3. E-commerce site crawling

[0071] The server crawls multiple e-commerce platforms and collects product data.

[0072] 4. Filtering and Ranking Results

[0073] The server filters the collected product data and ranks it based on criteria such as price and ratings.

[0074] 5. Search result presentation

[0075] The server sends the filtered and ranked results to the device.

[0076] The device displays the appropriate product list to the user.

[0077] 6. Personalized Offers

[0078] The server analyzes user behavior and makes more relevant personalized offers in the future.

[0079] In this way, a system is provided that allows users to easily find the product they are looking for, even if they do not know the specific product name or detailed conditions.

[0080] The processing flow will be explained below.

[0081] Step 1:

[0082] User: Enters vague criteria such as "I want autographed merchandise of a famous athlete" into the device's search bar. Once entered, presses the "Search" button.

[0083] Step 2:

[0084] Terminal: Receives the search criteria entered by the user, converts the data into an appropriate format (e.g., JSON), and sends it to the server. An example of sending data is {"query":"I want autographed merchandise by a famous athlete"}.

[0085] Step 3:

[0086] Server: Receives search criteria data sent from the device and temporarily stores this data in memory.

[0087] Step 4:

[0088] Server: Passes the received data to the generative AI model and begins the analysis process. The generative AI model analyzes the user's vague search criteria and converts them into an appropriate search query (e.g., "famous athlete signed merchandise").

[0089] Step 5:

[0090] Server: Stores the search queries generated by the generative AI model and prepares them to be sent to multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten).

[0091] Step 6:

[0092] Server: Uses generated search queries to access APIs and crawl the web for each e-commerce platform, and collects product data (price, description, ratings, etc.) from each platform.

[0093] Step 7:

[0094] Each e-commerce site: Returns a list of products based on the search query to the server through the provided API or web interface. For example, the results are returned in the format {"site": "Amazon", "results": [...]}.

[0095] Step 8:

[0096] Server: Aggregates product data collected from each e-commerce platform to form a large product dataset, which is temporarily stored in memory.

[0097] Step 9:

[0098] Server: Filters the aggregated product data. Based on the user's specified criteria (price range, ratings, etc.), the process narrows down the results to only those products that match the criteria.

[0099] Step 10:

[0100] Server: Ranks the filtered products. Ranking criteria include price, popularity, review rating, etc. Generates a ranked product list.

[0101] Step 11:

[0102] Server: Sends the filtered and ranked product list to the device. Example data sent includes {"ranked_products": [...]}.

[0103] Step 12:

[0104] Device: The received product list is displayed to the user in a format appropriate for the user interface, such as list view or grid view.

[0105] Step 13:

[0106] Users can select the product they are interested in from the list of products presented and proceed to the detail page or purchase page. If they do not like the product, they can enter new search criteria and search again.

[0107] Step 14:

[0108] Server: Records user actions (selection, confirmation of details, purchase, etc.) and stores them as analytical data to improve search accuracy next time.

[0109] Step 15:

[0110] Server: Prepares to make personalized product suggestions based on the analyzed user data.

[0111] This series of steps allows users to easily find the products they are looking for and allows the system to continuously improve its accuracy.

[0112] Example 1

[0113] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0114] When searching for a wide variety of products on the Internet, modern consumers often spend a lot of time and effort trying to find the product they are looking for based on vague or abstract criteria. Furthermore, existing search engines and e-commerce platforms are unable to fully utilize users' past search history and purchasing behavior, resulting in a lack of personalized and appropriate product suggestions. This makes it difficult for users to efficiently find the products they need.

[0115] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0116] In this invention, the server includes means for receiving vague conditions entered by a user and transmitting them to the server, means for the server to analyze the received vague conditions based on a generative AI model and convert them into an appropriate search query, means for the server to crawl multiple e-commerce platforms and aggregate product data based on the converted search query, means for the server to filter the aggregated product data and rank it based on the user's specified conditions, means for the server to transmit the filtered and ranked product data to a terminal and for the terminal to display it to the user, and means for the terminal to present the received product data to the user in a display format such as a list view or grid view, thereby enabling users to efficiently find desired products even with vague conditions.

[0117] A "user" is a person who enters vague criteria and receives search results.

[0118] "Device" means a device used by a User to enter search criteria and view product data returned by the Server. Examples include smartphones, tablets, and PCs.

[0119] The "server" refers to a central computer that receives vague conditions sent by users, converts them into search queries using a generative AI model, and retrieves the necessary product data from multiple e-commerce platforms.

[0120] A "generative AI model" refers to an artificial intelligence technique that analyzes vague conditions and converts them into specific search queries.

[0121] A "search query" refers to a search condition that converts a user's vague conditions into a specific and optimal form.

[0122] An "e-commerce platform" refers to an online marketplace that offers goods and services over the web. Examples include Amazon, eBay, and Rakuten.

[0123] "Product Data" refers to information retrieved from e-commerce platforms based on a search query, including prices, descriptions, user reviews, and ratings.

[0124] "Filtering" refers to the process of narrowing down the acquired product data based on the user's specified conditions.

[0125] "Ranking" refers to the process of hierarchically arranging filtered product data based on criteria such as price, popularity, and review rating.

[0126] "List view" refers to a method of displaying product data in a vertical list format.

[0127] "Grid view" refers to a method of displaying product data in multiple rows and columns.

[0128] A "feedback loop" is a system that analyzes users' search results and purchasing behavior history and uses them to improve the accuracy of their next search.

[0129] In an embodiment of the present invention, a system is provided that allows users to easily search for desired products and obtain optimal results. This system is composed of a series of processes: acquiring user input, analyzing conditions, crawling e-commerce sites, filtering and ranking the results, presenting optimal search results, and providing personalized suggestions and a feedback loop.

[0130] Hardware and Software Configuration

[0131] User

[0132] Users enter vague criteria into the search bar on their devices (smartphones, tablets, PCs, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[0133] Terminal

[0134] The device receives the information entered by the user. It then converts the data into an appropriate data format, such as JSON, and sends it to the server. For example, the vague query entered by the user, "I want autographed merchandise by a famous athlete," can be converted to "{"query": "I want autographed merchandise by a famous athlete"}".

[0135] server

[0136] The server receives vague conditions sent from the device. This data is passed to a generative AI model (e.g., GPT-3). The generative AI model analyzes the user's vague conditions and converts them into specific search queries (e.g., "famous athlete signed merchandise"). The converted search queries are optimized for crawling e-commerce sites (e.g., Amazon, eBay, Rakuten).

[0137] The server crawls multiple e-commerce platforms using the search query generated by the generative AI model, sends the search query to each platform, and retrieves relevant product data, including prices, descriptions, user reviews, and ratings.

[0138] The server then aggregates the product data obtained from each e-commerce platform. This aggregated product data is first filtered and narrowed down based on user-specified criteria (e.g., price range, rating, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review rating.

[0139] The ranked product data is sent to the device, which then presents the received product data to the user in a display format such as list view or grid view, allowing users to easily find the best product from a large number of options.

[0140] Finally, the server analyzes the user's past purchasing behavior and search history to provide personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to refine the next search criteria. This allows the user to receive more personalized search results and product suggestions.

[0141] Specific operation example

[0142] For example, if a user wants to find autographed merchandise of a famous athlete and enters the search criteria "I want autographed merchandise of a famous athlete," the system will perform the following actions:

[0143] 1. Getting User Input

[0144] The user types "I want autographed merchandise of a famous athlete" into their device and clicks the "Search" button.

[0145] 2. Analysis of Conditions

[0146] The server passes the received input to a generative AI model, which generates a specific search query: "famous athlete signed merchandise."

[0147] 3. Crawling e-commerce sites

[0148] The server uses the generated search query to crawl multiple e-commerce platforms, such as Amazon, eBay, and Rakuten, to collect relevant product data.

[0149] 4. Filtering and Ranking Results

[0150] The server filters the collected product data based on price, user reviews, etc., and ranks them.

[0151] 5. Providing optimal search results

[0152] The server sends the filtered and ranked product data to the device, which then displays it to the user in a list view, showing each product's image, price, and rating.

[0153] 6. Personalized recommendations and feedback loops

[0154] The server analyzes the products the user clicks on and their purchase history, allowing it to suggest more suitable products for the next search.

[0155] Prompt Sentence Examples

[0156] As an example of a prompt statement, enter the following:

[0157] If a user types in "I want autographed merchandise of a famous athlete," analyze what kind of search query this translates into.

[0158] As described above, the present invention realizes a system that enables a user to efficiently find a desired product even if the user has vague conditions.

[0159] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0160] Step 1:

[0161] A user uses their device to enter vague search criteria into the search bar, such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[0162] Specifically, the user types text into an input field and performs a submit operation.

[0163] Input: Vague conditions entered by the user (e.g., "I want autographed merchandise by a famous athlete")

[0164] Output: Data when the send button is pressed and the search conditions are confirmed

[0165] Step 2:

[0166] After receiving the information entered by the user, the device converts the data into an appropriate data format, such as JSON, and sends it to the server. For example, it converts the data into JSON data such as "{'query': 'I want autographed merchandise of a famous athlete'}".

[0167] Specifically, the device converts the input text into a JSON string and sends it to the server as an HTTP request.

[0168] Input: User input data (e.g., "I want autographed merchandise from a famous athlete")

[0169] Output: Search criteria data in JSON format (e.g., "{'query': 'I want autographed merchandise by famous athletes'}")

[0170] Step 3:

[0171] The server receives the vague search terms sent from the device and passes this data to the generative AI model, which then analyzes the user's vague search terms and converts them into a specific search query (e.g., "famous athlete signed merchandise").

[0172] Specifically, the server calls an API endpoint and passes the data to the generative AI model for analysis.

[0173] Input: Search criteria data in JSON format (e.g., "{'query': 'I want autographed merchandise by famous athletes'}")

[0174] Output: A specific search query (e.g., "famous athlete signed merchandise")

[0175] Step 4:

[0176] The server crawls multiple e-commerce platforms using the search query generated by the generative AI model, sending the search query to each platform and retrieving relevant product data (e.g., price, description, user reviews, ratings, etc.).

[0177] Specifically, the server runs a crawling script and sends API requests to each platform to collect product data.

[0178] Input: A specific search query (e.g., "famous athlete signed merchandise")

[0179] Output: Product data obtained from each platform (e.g. price, description, user reviews, ratings)

[0180] Step 5:

[0181] The server aggregates product data obtained from each e-commerce platform. The aggregated product data is then filtered and narrowed down based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[0182] Specifically, the server executes code to sort and filter the product data.

[0183] Input: Raw data from each platform (e.g. price, description, user reviews, ratings)

[0184] Output: Filtered and ranked product data

[0185] Step 6:

[0186] The server transmits the filtered and ranked product data to the terminal, which presents the received product data to the user in a display format such as a list view or a grid view.

[0187] Specifically, the product list configured on the server side is sent to the terminal as an HTTP response, and the data is displayed as a UI on the terminal side.

[0188] Input: Filtered and ranked product data

[0189] Output: A list of products displayed in list view or grid view

[0190] Step 7:

[0191] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to improve the accuracy of the next search criteria. This provides the user with more personalized search results and product suggestions.

[0192] Specifically, the server accesses the user database and runs an algorithm that analyzes past data.

[0193] Input: User's past purchasing behavior and search history

[0194] Output: personalized product recommendations

[0195] (Application example 1)

[0196] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0197] Conventional e-commerce systems have the problem that it is difficult for users to find the desired product unless they enter precise keywords. Furthermore, searches using vague conditions often result in the display of many irrelevant products, making them inefficient for users. Meanwhile, individual suggestions based on purchase and search history are not fully utilized, making it difficult to suggest optimal products tailored to the user's needs. Furthermore, the limited options for displayed products and the lack of feedback reflected in the next search also detract from the user experience.

[0198] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0199] In this invention, the server includes a means for receiving vague search terms entered by a user, analyzing them based on a generative AI model, and converting them into an appropriate search query; a means for crawling multiple e-commerce platforms and aggregating product data based on the converted search query; and a means for filtering the aggregated product data and ranking them based on the user's specified search terms. This allows users to search for optimal products even with vague search terms. The server also uses the generative AI model to analyze past purchasing behavior and make personalized product suggestions, thereby providing more appropriate products to users. Additionally, the system includes a means for the terminal to allow users to select additional products based on the displayed search results, and a means for sending user feedback to the server to create a feedback loop to improve the accuracy of the next search, thereby improving the user experience.

[0200] "User input" refers to the action of a user providing their desired conditions or requests through an electronic device.

[0201] A "generative AI model" is an algorithm or system that uses artificial intelligence technology to convert vague conditions into specific search queries.

[0202] A "search query" is a keyword or phrase that a user enters into a search engine to retrieve specific information.

[0203] "E-commerce platform" refers to a website or application for buying and selling products and services over the Internet.

[0204] "Crawling" is the process of automatically visiting web pages on the Internet and collecting data.

[0205] "Product data" is a data set that contains detailed product information, such as price, description, and reviews.

[0206] "Filtering" is a process for narrowing down data based on specific conditions.

[0207] "Ranking" is the process of ranking data based on specific criteria.

[0208] "Personalized suggestions" analyze a user's past behavior and preferences to suggest products and services that are best suited to each individual user.

[0209] A "feedback loop" is a system that collects user reactions and behavior and reflects them in the next service provision.

[0210] The system for realizing this invention mainly includes a terminal that processes user input, a server that analyzes ambiguous conditions and generates optimal search queries, and a crawling function that aggregates product data from e-commerce platforms based on the generated queries.

[0211] User input via terminal

[0212] The user inputs their desired conditions and requests into the terminal. For example, the user might input "I want autographed merchandise of a famous athlete." Once the input is confirmed, the terminal sends this information to the server. The terminal can be a smartphone, tablet, PC, or other device.

[0213] Server-based analysis of ambiguous conditions

[0214] The server passes the vague conditions received from the device to a generative AI model. For example, the BART transformer model is used as the generative AI model. The model converts the vague conditions into a specific search query. For example, the condition "I want signed merchandise by a famous athlete" is converted to "famous athlete signed merchandise."

[0215] Crawling product data from e-commerce sites

[0216] The server uses the generated search query to crawl multiple e-commerce platforms (e.g., Amazon and Rakuten) to retrieve relevant product data, including information such as price, description, reviews, and ratings.

[0217] Filtering and Ranking Data

[0218] The acquired product data is filtered on the server and ranked based on user-specified criteria (price range, ratings, etc.) The filtering and ranking algorithms are implemented in Python, for example.

[0219] Providing optimal search results

[0220] The server sends the filtered and ranked product data to the device, which displays it to the user, who can then browse the best products in a list or grid view and select specific products.

[0221] Personalized product recommendations

[0222] The server analyzes the user's past purchasing behavior and search history and uses a generative AI model to make personalized suggestions, thereby providing users with more relevant and personalized product suggestions.

[0223] Formation of a feedback loop

[0224] The device sends user operation information and feedback to the server, which then analyzes the feedback and creates a feedback loop to improve the accuracy of the next search.

[0225] Specific examples

[0226] For example, if a user types in "latest smartphone case recommendations," the server converts this to "latest smartphone cases recommendations," retrieves smartphone case information from Amazon and Rakuten, and displays a list.

[0227] Prompt Sentence Examples

[0228] "Generate what query would be converted when a user types 'I want autographed merchandise of a famous athlete' into a product search app."

[0229] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0230] Step 1:

[0231] Get the conditions entered by the user

[0232] The device accepts input from the user. Specifically, the user enters vague keywords such as "I want autographed merchandise of a famous athlete" into the product search bar. Once this input is confirmed, the device converts the information into a data format such as JSON and sends the data to the server.

[0233] Input: Vague keywords entered by the user

[0234] Output: Keyword information converted into data format such as JSON

[0235] Step 2:

[0236] Convert vague conditions into specific search queries

[0237] The server passes the vague conditions received from the device to a generative AI model. The generative AI model (e.g., the BART transformer model) analyzes the conditions and converts them into specific search queries. For example, "I want autographed merchandise by a famous athlete" is converted to "famous athlete signed merchandise."

[0238] Input: Vague keywords in JSON format sent from the terminal

[0239] Output: A specific search query transformed by the generative AI model

[0240] Step 3:

[0241] Crawl multiple e-commerce platforms based on a search query

[0242] The server uses the generated search query to access the APIs of multiple e-commerce platforms (e.g., Amazon and Rakuten) and crawl product data, including product names, prices, descriptions, reviews, ratings, etc.

[0243] Input: A specific search query transformed by a generative AI model

[0244] Output: Product data obtained from multiple e-commerce platforms

[0245] Step 4:

[0246] Filter and rank your product data

[0247] The server aggregates the acquired product data and filters it based on the user's specified conditions (price range, ratings, etc.), and then ranks the filtered product data based on criteria such as ratings and price.

[0248] Input: Product data obtained from multiple e-commerce platforms

[0249] Output: Filtered and ranked product data

[0250] Step 5:

[0251] Displaying the most relevant product information to users

[0252] The server transmits the filtered and ranked product data to the terminal, which displays the data to the user in a format such as a list view or a grid view, allowing the user to review and select products.

[0253] Input: Filtered and ranked product data

[0254] Output: Product information displayed in list or grid view format

[0255] Step 6:

[0256] Make personalized product recommendations based on past purchasing behavior and search history

[0257] The server analyzes the user's past purchasing behavior and search history, and uses a generative AI model based on that information to make personalized product suggestions, allowing users to be recommended products that match their preferences.

[0258] Input: User's past purchasing behavior and search history

[0259] Output: personalized product recommendations

[0260] Step 7:

[0261] Forming a feedback loop

[0262] The device sends user operation information and feedback to the server, which analyzes this feedback and reflects it in order to improve the accuracy of the next search, allowing the user to obtain more accurate search results.

[0263] Input: User operation information and feedback

[0264] Output: Data reflected to improve search accuracy next time

[0265] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0266] In an embodiment of the present invention, a system is provided that allows users to easily search for desired products using vague search criteria and obtain optimal results. Furthermore, the system recognizes the user's emotions and improves the accuracy of search queries based on those emotions, thereby providing more personalized product offers. A specific embodiment of the system and its operation are described below.

[0267] Getting User Input

[0268] User

[0269] Users enter vague criteria into the search bar on their device (smartphone, tablet, PC, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." Once they've finished entering the information, they click the "Search" button.

[0270] Terminal

[0271] The terminal receives the information entered by the user, converts the data into an appropriate data format such as JSON, and sends it to the server.

[0272] Condition analysis and emotion recognition

[0273] server

[0274] The server receives the ambiguous condition data sent from the device, passes this data to the generative AI model, and begins the analysis process.

[0275] Generative AI Models

[0276] The generative AI model analyzes the user's vague search criteria and converts them into a specific search query, which is then used to provide optimal search results.

[0277] Emotion Engine

[0278] The emotion engine analyzes the user's emotions at the time of input from voice, text input, facial expressions, etc. Based on the results of this analysis, the generative AI model optimizes its condition analysis. For example, if the user is excited, it will make more proactive product suggestions, while if the user is calm, it will emphasize more detailed information.

[0279] Crawling e-commerce sites

[0280] server

[0281] The server uses the generated search query and the sentiment data obtained from the sentiment engine to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.) and sends a search query to each platform to retrieve related product data.

[0282] Filtering and ranking results

[0283] server

[0284] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. This dataset is then filtered to narrow down the list based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[0285] Providing optimal search results

[0286] server

[0287] The server transmits the filtered and ranked product data to the terminal.

[0288] Terminal

[0289] The device displays the received product data to the user in a list view, grid view, or other display format, allowing the user to easily find the best product from a wide range of options.

[0290] Personalized recommendations and feedback loops

[0291] server

[0292] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is stored and used to improve the accuracy of the next search criteria. User emotional data is also stored and reflected in future searches and suggestions.

[0293] user

[0294] Users can select a product from the product list displayed on their device and proceed to the details page or purchase page. They can also enter new search criteria and search again.

[0295] Specific examples

[0296] For example, suppose a user who wants to find autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete" and seems to be having fun. In this case, the system will perform the following actions.

[0297] 1. User Input

[0298] The user types "I want autographed merchandise from a famous athlete" into the device.

[0299] 2. Condition Analysis and Emotion Recognition

[0300] The server passes the input to a generative AI model for analysis, resulting in the search query "famous athlete signed merchandise."

[0301] The emotion engine analyzes the user's positive emotions and, based on the generated query, prioritizes products with more positive reviews.

[0302] 3. E-commerce site crawling

[0303] The server crawls multiple e-commerce platforms and collects product data.

[0304] 4. Filtering and Ranking Results

[0305] The server filters the collected product data and prioritizes and ranks products that have high ratings and many positive reviews that reflect the user's enjoyment emotions.

[0306] 5. Search result presentation

[0307] The server sends the filtered and ranked results to the device.

[0308] The device displays the appropriate product list to the user.

[0309] 6. Personalized Offers

[0310] The server analyzes the user's behavioral and emotional data to make more relevant personalized suggestions in the future.

[0311] In this way, a system is provided that allows users to easily find the products they are looking for without knowing the specific product names or detailed conditions, and furthermore, allows them to receive appropriate product suggestions based on their emotions.

[0312] The processing flow will be explained below.

[0313] Step 1:

[0314] User: Enters vague criteria such as "I want autographed merchandise of a famous athlete" into the device's search bar. Once entered, presses the "Search" button.

[0315] Step 2:

[0316] Terminal: Receives the search criteria entered by the user, converts the data into an appropriate format (e.g., JSON), and sends it to the server. An example of sending data is {"query":"I want autographed merchandise by a famous athlete"}.

[0317] Step 3:

[0318] Server: Receives search criteria data sent from the device and temporarily stores this data in memory.

[0319] Step 4:

[0320] Server: Passes the received data to the generative AI model and begins the analysis process. The generative AI model analyzes the user's vague search criteria and converts them into an appropriate search query (e.g., "famous athlete signed merchandise").

[0321] Step 5:

[0322] Emotion engine: Analyzes the emotions at the time of input from the user's voice, text input, facial expressions, etc. The emotion engine outputs the analysis results and optimizes the condition analysis of the generative AI model.

[0323] Step 6:

[0324] Server: Stores the search queries generated by the generative AI model and the sentiment analysis data from the sentiment engine, and prepares them to be sent to multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.).

[0325] Step 7:

[0326] Server: Uses generated search queries to access APIs and crawl the web for each e-commerce platform, and collects product data (price, description, ratings, etc.) from each platform.

[0327] Step 8:

[0328] Each e-commerce site: Returns a list of products based on the search query to the server through the provided API or web interface. For example, the results are returned in the format {"site": "Amazon", "results": [...]}.

[0329] Step 9:

[0330] Server: Aggregates product data collected from each e-commerce platform to form a large product dataset, which is temporarily stored in memory.

[0331] Step 10:

[0332] Server: Filters the aggregated product data. Based on the user's criteria (price range, ratings, etc.), it runs a process to narrow down the list to only products that match the criteria. It also prioritizes products based on sentiment data (e.g., if the user seems happy, prioritize products with many positive reviews).

[0333] Step 11:

[0334] Server: Ranks the filtered products. Ranking criteria include price, popularity, review rating, etc. Generates a ranked product list.

[0335] Step 12:

[0336] Server: Sends the filtered and ranked product list to the device. Example data sent includes {"ranked_products": [...]}.

[0337] Step 13:

[0338] Device: The received product list is displayed to the user in a format appropriate for the user interface, such as list view or grid view.

[0339] Step 14:

[0340] Users can select the product they are interested in from the list of products presented and proceed to the detail page or purchase page. If they do not like the product, they can enter new search criteria and search again.

[0341] Step 15:

[0342] Server: Records user behavior (selection, confirmation of details, purchase, etc.) and emotional data, and stores it as analytical data to improve search accuracy next time.

[0343] Step 16:

[0344] Server: Prepares to make future personalized product recommendations based on the analyzed user and sentiment data.

[0345] This series of steps allows users to easily find the products they are looking for and allows the system to continuously improve its accuracy.

[0346] Example 2

[0347] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0348] In conventional e-commerce platforms, when users search for products using vague criteria, they often encounter problems such as insufficient search results or being unable to find the desired product. Furthermore, because the platform does not take into account the user's purchasing motivation or emotions, personalized product suggestions are not possible, often resulting in low satisfaction. The present invention aims to solve these problems and provide users with optimal product search results and personalized suggestions.

[0349] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0350] In this invention, the server includes a means for analyzing vague search terms entered by a user based on a generative AI model and converting them into an appropriate search query, an emotion recognition means for analyzing the user's emotional state, and a means for crawling multiple e-commerce platforms and aggregating product data based on the converted search query. This allows a specific search query appropriate to the user's intent to be generated, even if the user enters vague search terms, and the aggregated product data can be ranked based on the user's specified search terms and emotional state. Furthermore, the system analyzes the user's past behavioral data and emotional data to provide personalized product suggestions, thereby increasing user satisfaction.

[0351] "Ambiguous search conditions entered by the user" refers to non-specific search conditions entered by the user through the terminal.

[0352] A "generative AI model" refers to an artificial intelligence model that analyzes vague conditions entered by users and converts them into appropriate search queries.

[0353] A "search query" refers to a specific search term or phrase that a generative AI model generates by analyzing ambiguous conditions.

[0354] "Emotion recognition means" refers to technology that analyzes the emotional state of a user from their input or interaction, including, for example, voice, text input, and facial expression recognition.

[0355] An "e-commerce platform" is a website or application that enables the buying and selling of goods and services online, such as Amazon or eBay.

[0356] "Crawling" refers to the act of a server automatically searching for information on a particular website and collecting data.

[0357] "Product Data" refers to all product-related information obtained from the e-commerce platform, such as product name, price, review rating, etc.

[0358] "Filtering" refers to the process of narrowing down collected product data based on user-specified conditions.

[0359] "Ranking" refers to the process of ranking filtered product data based on specific criteria, such as price, popularity, or review rating.

[0360] "Specified conditions" refer to conditions that a user sets to narrow down search results, such as a price range or rating range.

[0361] "Individualized product proposals" refer to product proposals that are optimized for each individual user based on the user's past behavioral and emotional data.

[0362] A "feedback loop" refers to a system that receives feedback from users and uses that information to improve the accuracy of subsequent searches and suggestions.

[0363] The present invention is a system for providing optimal search results when a user searches for products under vague conditions. Furthermore, the system recognizes the user's emotions and improves the accuracy of the search query based on the emotions, thereby providing more personalized product suggestions. A specific embodiment of the system and its operation are described below.

[0364] 1. Getting User Input

[0365] User

[0366] Users enter vague criteria into the search bar on their device (smartphone, tablet, PC, etc.), for example, an abstract request such as "I want autographed merchandise of a famous athlete." Once they're done, they click the "Search" button.

[0367] Terminal

[0368] The device acquires the conditions entered by the user, converts this input data into an appropriate data format such as JSON, and sends it to the server. At this time, when the user has completed entering the information on the device, the device sends an HTTP request to the server.

[0369] 2. Condition analysis and emotion recognition

[0370] server

[0371] The server receives vague condition data sent from the device. This data is passed to the generative AI model as a prompt. For example, the data {"query": "I want autographed merchandise of a famous athlete"} is sent.

[0372] Generative AI Models

[0373] The generative AI model analyzes the received prompt and converts the user's vague search criteria into a specific search query, such as "famous athlete signed merchandise." This converted query is cached on the server and can be reused.

[0374] Emotion Engine

[0375] The emotion engine analyzes user input (voice, text, facial expressions, etc.) to recognize the user's emotional state. For example, facial recognition technology can be used to determine whether a user is happy when typing. This emotional data influences the analysis of the generative AI model.

[0376] 3. Crawling e-commerce sites

[0377] server

[0378] The server uses the converted search query and sentiment data to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.). For each platform, it sends a new search query and collects related product data. For example, it uses the Amazon API to send the query "famous athlete signed merchandise" and retrieves the corresponding product data.

[0379] 4. Filtering and Ranking Results

[0380] server

[0381] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. This dataset is then filtered to narrow down the results based on user-specified criteria (price range, ratings, etc.). Specifically, it selects highly rated products within a certain price range. It also prioritizes products with proactive product suggestions and detailed information based on emotional data. For example, for users with a positive emotional state, it prioritizes products with many positive reviews.

[0382] 5. Providing optimal search results

[0383] server

[0384] The server sends the filtered and ranked product data to the terminal, for example, returning a product list in JSON format.

[0385] Terminal

[0386] The device displays the received product data to the user. The display format can be selected from list view, grid view, etc. For example, in list view, the product name, image, price, and review rating are displayed.

[0387] 6. Personalized recommendations and feedback loops

[0388] server

[0389] The server analyzes users' past purchasing behavior and search history to make personalized product suggestions. It also accumulates behavioral data, such as when users click on a specific product to view its details or make a purchase. This data is analyzed and saved along with emotional data to help improve the accuracy of the next search.

[0390] User

[0391] Users can select products they are interested in from the products displayed on their devices, proceed to the details and purchase process, or enter new search criteria and search again.

[0392] Specific examples

[0393] For example, suppose a user searching for autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete" and seems to be having fun. In this case, the system performs the following specific actions:

[0394] 1. User Input

[0395] The user types "I want autographed merchandise from a famous athlete" into the device.

[0396] The user clicks the Search button.

[0397] 2. Condition Analysis and Emotion Recognition

[0398] The server passes the input data to the generative AI model.

[0399] A generative AI model generates the search query "famous athlete signed merchandise."

[0400] The emotion engine analyzes the user's happy facial expression.

[0401] 3. E-commerce site crawling

[0402] A server sends search queries to multiple e-commerce platforms to collect product data.

[0403] 4. Filtering and Ranking Results

[0404] The server filters the collected product data and prioritizes products with positive reviews and high ratings.

[0405] 5. Search result presentation

[0406] The server transmits the ranked product data to the terminal.

[0407] The device displays the product list to the user in a list view format.

[0408] 6. Personalized Offers

[0409] The server analyzes the user's behavioral and emotional data to make more appropriate personalized suggestions in the future.

[0410] Example prompt sentence:

[0411] "I want autographed merchandise from famous athletes."

[0412] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0413] Step 1:

[0414] Getting User Input

[0415] The user enters vague search criteria into the search bar of their device (smartphone, tablet, PC, etc.). For example, they might type "I want autographed merchandise of a famous athlete." Then, they click the "Search" button. The input data is in text format.

[0416] Step 2:

[0417] Sending data from the device to the server

[0418] The device converts the user-entered criteria into an appropriate data format, such as JSON. The device then sends this data to the server using an HTTP request, such as {"query": "I want autographed merchandise by a famous athlete"}. This data is provided by the user as input.

[0419] Step 3:

[0420] Condition Analysis

[0421] The server passes the vague condition data received from the device to the generative AI model. The generative AI model analyzes this data and converts it into a specific search query. For example, the vague condition "I want signed merchandise by a famous athlete" is converted into the specific search query "famous athlete signed merchandise." This converted query is output as search data.

[0422] Step 4:

[0423] Emotion Analysis

[0424] The server uses an emotion engine to analyze the user's input format (voice, text, facial expression, etc.). The emotion engine analyzes emotions from the user's facial expression and voice data at the time of input and returns the results to the server. For example, facial expression analysis can yield an emotion result of "looking happy." This emotion data becomes input for use in optimizing search queries.

[0425] Step 5:

[0426] Crawling e-commerce sites

[0427] The server uses the generated search query and the analyzed sentiment data to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.). During the crawling, it uses the platform's API or scraping technology to send the search query and retrieve the corresponding product data. For example, it sends the search query "famous athlete signed merchandise" to Amazon to retrieve a list of related products. This product data is output as a result from the e-commerce site.

[0428] Step 6:

[0429] Filtering and ranking results

[0430] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. Products are filtered from this dataset based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on emotion data. For example, products with many positive reviews are prioritized for users with positive emotions. The filtered and ranked product data is output as an optimal product list.

[0431] Step 7:

[0432] Providing optimal search results

[0433] The server sends the filtered and ranked product data to the device, for example in the format {"products": [listed product data]}. This data is returned to the user as the best search results.

[0434] Terminal

[0435] The device displays the received product data to the user. The display format can be selected from list view, grid view, etc. For example, list view displays the product name, image, price, and review rating. This displayed product information is provided in a format that is easy for the user to view.

[0436] Step 8:

[0437] Personalized recommendations and feedback loops

[0438] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. It also accumulates behavioral data, such as when a user clicks on a specific product to view its details page or make a purchase. This data is analyzed to help improve the accuracy of the next search criteria. For example, if a user repeatedly takes positive action regarding "signature goods," similar products will be included in the next suggestion. This feedback data is output as the user's behavioral pattern.

[0439] User

[0440] Users can select products they are interested in from the displayed products and proceed to the purchase process. They can also search again by entering new search criteria, improving the user experience. These new search criteria are then processed as data to be input again into the system.

[0441] As described above, the system of the present invention can efficiently carry out a series of processes from vague conditions to specific product proposals.

[0442] (Application example 2)

[0443] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0444] Existing search systems have difficulty finding appropriate search results for vague user inputs, making it difficult for users to smoothly find the products they are looking for. Furthermore, the lack of personalized suggestions based on user emotions and past behavior makes it difficult to provide a more personalized shopping experience. To solve these problems, there is a need for a search system that takes user emotions into account.

[0445] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving ambiguous conditions entered by a user and transmitting them to the server, means for the server to analyze the ambiguous conditions received based on a generative AI model and convert them into an appropriate search query, means for the server to crawl multiple e-commerce platforms and aggregate product data based on the converted search query, means for the server to filter the aggregated product data using an emotion recognition engine taking into account the user's emotional data and rank it based on the user's specified conditions, and means for the server to transmit the filtered and ranked product data to a terminal, which then displays it to the user. This enables users to accurately search for products even with ambiguous conditions and receive more personalized product suggestions based on their emotions.

[0446] "User" means an individual or organization that uses the System to search for and purchase products.

[0447] "Vague terms" are abstract search queries or requests entered by users without specifying specific products or specifications.

[0448] "Server" refers to a computer system that receives input data from users and performs processes such as analysis using generative AI models, emotion recognition engines, crawling e-commerce platforms, filtering and ranking data, etc.

[0449] A "generative AI model" is a type of artificial intelligence that analyzes ambiguous conditions and converts them into appropriate search queries, and is based on natural language processing.

[0450] An "emotion recognition engine" is a technology for analyzing emotions from non-verbal information such as a user's facial expressions and voice.

[0451] An "e-commerce platform" is an online marketplace for buying and selling goods and services over the Internet.

[0452] "Crawling" is the process of automatically gathering information on a website.

[0453] "Filtering" is the process of removing unnecessary data from collected product data based on specific criteria and extracting only the necessary data.

[0454] "Ranking" is the process of ranking the filtered product data based on specific evaluation criteria.

[0455] A "terminal" is a device that a user uses to access the system, including smartphones and computers.

[0456] A "feedback loop" is a cyclical process that analyzes information collected from users (e.g., search history, sentiment data) to improve the accuracy of future search criteria and suggestions.

[0457] The system for implementing this invention allows users to search for products under vague conditions and optimizes the results based on emotions. The specific configuration and operation of this system will be described below.

[0458] System Configuration

[0459] Hardware

[0460] Smartphone: A device with a built-in camera and microphone that allows users to access the system.

[0461] Server: A computer system equipped with a high-speed processor and capable of processing large amounts of data. Cloud-based operation is recommended.

[0462] software

[0463] Generative AI model: A natural language processing model to convert users' vague search terms into specific search queries. In this case, we use OpenAI's GPT-4.

[0464] Emotion recognition engine: An engine that analyzes the user's facial expressions and voice to recognize emotions. It uses Google Cloud Vision API and Microsoft Azure Face API.

[0465] E-commerce site crawling library: A library for collecting product data from e-commerce platforms. It uses BeautifulSoup (Python) and other tools.

[0466] Cloud platform: Infrastructure for operations and data processing. We recommend using Amazon Web Services (AWS).

[0467] System Operation

[0468] 1. Getting User Input

[0469] Users enter vague terms into the search bar on their smartphone and press the search button.

[0470] Example: "I want new shoes."

[0471] 2. Emotion recognition

[0472] The device's camera and microphone are used to collect the user's facial expressions and voice.

[0473] The emotion recognition engine analyzes this data and recognizes the user's emotions.

[0474] Example: A user looks happy.

[0475] 3. Condition analysis

[0476] Vague search queries are fed into a generative AI model, which converts them into specific search queries.

[0477] Example: "new shoes"

[0478] 4. Crawling e-commerce sites

[0479] The server uses the generated search query to crawl multiple e-commerce platforms to collect product data.

[0480] The collected data is converted into a unified format and aggregated on the server.

[0481] 5. Filtering and Ranking Results

[0482] The aggregated product data is filtered taking into account user sentiment data.

[0483] The filtered data is then ranked based on criteria such as price, popularity, and review ratings.

[0484] 6. Presentation of search results

[0485] The server transmits the filtered and ranked data to the terminal.

[0486] The terminal displays a list of products that are most suitable for the user.

[0487] 7. Creating feedback loops

[0488] User selections, search history, emotional data, etc. are stored on the server to improve search accuracy from the next time onwards.

[0489] Prompt Sentence Examples

[0490] Here is an example of a prompt for the generative AI model:

[0491] A user searched for: "I want new shoes." The user's sentiment seems to be happy. Convert this into a specific search query.

[0492] In this way, users can easily find the products they are looking for without having to enter specific product names or detailed search criteria. Furthermore, the system's emotion-based product suggestions enable a more personalized shopping experience.

[0493] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0494] Step 1: Getting User Input

[0495] The user enters vague search criteria into the device, for example, "I want new shoes." Once the search is complete, the user presses the search button. The device converts the input data into JSON format and sends it to the server.

[0496] Input: User's vague search criteria

[0497] Output: Input data in JSON format

[0498] Specific actions: Enter search criteria in the search bar and press the button.

[0499] Step 2: Emotion Recognition

[0500] The device uses a built-in camera and microphone to collect the user's facial expressions and voice. This data is sent to an emotion recognition engine to analyze the user's emotions. The analysis results are sent to the server in JSON format.

[0501] Input: User's facial expression and voice data

[0502] Output: Emotion data in JSON format

[0503] Specific actions: Capture facial expressions and voice using a camera and microphone.

[0504] Step 3: Condition analysis

[0505] The server receives the search criteria and emotion data sent from the device, passes this data to the generative AI model, and converts it into a specific search query. The generative AI model analyzes the prompt and generates an appropriate search query.

[0506] Input: Fuzzy search terms and sentiment data in JSON format

[0507] Output: Specific search query

[0508] Specific operation: The generative AI model analyzes the prompt sentence and generates a search query.

[0509] Step 4: Crawl your e-commerce site

[0510] The server crawls multiple e-commerce platforms using the generated specific search query, collects product data using libraries such as BeautifulSoup, converts the collected data into a unified format, and aggregates it on the server.

[0511] Input: A specific search query

[0512] Output: Aggregated product data in a unified format

[0513] Specific operation: Collects product data using e-commerce site crawl libraries.

[0514] Step 5: Filter and rank the results

[0515] The server receives the aggregated product data and filters it using an emotion recognition engine, taking into account the user's emotional data. The filtered data is then ranked based on criteria such as price, ratings, and reviews.

[0516] Input: Aggregated product data and sentiment data in a unified format

[0517] Output: Filtered and ranked product data

[0518] Specific behavior: Analyze product data and optimize based on user sentiment.

[0519] Step 6: Presenting search results

[0520] The server sends the filtered and ranked product data to the device, which then visually displays this data to the user in a format such as a list view or grid view.

[0521] Input: Filtered and ranked product data

[0522] Output: The optimal product list shown to the user

[0523] Specific operation: Display the product list on the terminal screen.

[0524] Step 7: Creating a feedback loop

[0525] The server stores user behavioral history and emotional data to improve the accuracy of search criteria and suggestions from the next time onward, allowing users to receive more personalized search results.

[0526] Input: User behavior history, emotional data

[0527] Output: Feedback to improve accuracy using accumulated data

[0528] Specific actions: Accumulate behavioral history and emotional data and reflect them in the next proposal.

[0529] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0530] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0531] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0532] [Second embodiment]

[0533] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0534] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0535] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0536] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0537] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0538] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0539] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0540] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0541] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0542] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0543] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0544] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0545] In the embodiment of the present invention, a system is provided that allows users to easily search for products they want and obtain optimal results. A specific embodiment of this system and its operation will be described below.

[0546] Getting User Input

[0547] User

[0548] Users enter vague criteria into the search bar on their devices (smartphones, tablets, PCs, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[0549] Terminal

[0550] The terminal accepts information entered by the user, converts the data into an appropriate data format such as JSON, and sends it to the server.

[0551] Condition Analysis

[0552] server

[0553] The server receives vague conditions sent from the device. This data is passed to the generative AI model. The generative AI model analyzes the user's vague conditions and converts them into a specific search query. This converted search query is optimized to search for the product the user is looking for.

[0554] Crawling e-commerce sites

[0555] server

[0556] The server crawls multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.) using the search query generated by the generative AI model. It sends the search query to each platform and retrieves relevant product data, including prices, descriptions, user reviews, and ratings.

[0557] Filtering and ranking results

[0558] server

[0559] The server aggregates product data obtained from each e-commerce platform. The aggregated product data is filtered and narrowed down based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[0560] Providing optimal search results

[0561] server

[0562] The server transmits the filtered and ranked product data to the terminal.

[0563] Terminal

[0564] The device displays the received product data to the user in a list view, grid view, or other display format, allowing the user to easily find the best product from a wide range of options.

[0565] Personalized recommendations and feedback loops

[0566] server

[0567] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to improve the accuracy of the next search criteria. This provides the user with more personalized search results and product suggestions.

[0568] user

[0569] Users can select a product from the product list displayed on their device and proceed to the details page or purchase page. They can also enter new search criteria and search again.

[0570] Specific examples

[0571] For example, if a user who wants to find autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete," the system will perform the following actions.

[0572] 1. User Input

[0573] The user types "I want autographed merchandise from a famous athlete" into the device.

[0574] 2. Condition analysis

[0575] The server passes the input to a generative AI model for analysis, resulting in the search query "famous athlete signed merchandise."

[0576] 3. E-commerce site crawling

[0577] The server crawls multiple e-commerce platforms and collects product data.

[0578] 4. Filtering and Ranking Results

[0579] The server filters the collected product data and ranks it based on criteria such as price and ratings.

[0580] 5. Search result presentation

[0581] The server sends the filtered and ranked results to the device.

[0582] The device displays the appropriate product list to the user.

[0583] 6. Personalized Offers

[0584] The server analyzes user behavior and makes more relevant personalized offers in the future.

[0585] In this way, a system is provided that allows users to easily find the product they are looking for, even if they do not know the specific product name or detailed conditions.

[0586] The processing flow will be explained below.

[0587] Step 1:

[0588] User: Enters vague criteria such as "I want autographed merchandise of a famous athlete" into the device's search bar. Once entered, presses the "Search" button.

[0589] Step 2:

[0590] Terminal: Receives the search criteria entered by the user, converts the data into an appropriate format (e.g., JSON), and sends it to the server. An example of sending data is {"query":"I want autographed merchandise by a famous athlete"}.

[0591] Step 3:

[0592] Server: Receives search criteria data sent from the device and temporarily stores this data in memory.

[0593] Step 4:

[0594] Server: Passes the received data to the generative AI model and begins the analysis process. The generative AI model analyzes the user's vague search criteria and converts them into an appropriate search query (e.g., "famous athlete signed merchandise").

[0595] Step 5:

[0596] Server: Stores the search queries generated by the generative AI model and prepares them to be sent to multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten).

[0597] Step 6:

[0598] Server: Uses generated search queries to access APIs and crawl the web for each e-commerce platform, and collects product data (price, description, ratings, etc.) from each platform.

[0599] Step 7:

[0600] Each e-commerce site: Returns a list of products based on the search query to the server through the provided API or web interface. For example, the results are returned in the format {"site": "Amazon", "results": [...]}.

[0601] Step 8:

[0602] Server: Aggregates product data collected from each e-commerce platform to form a large product dataset, which is temporarily stored in memory.

[0603] Step 9:

[0604] Server: Filters the aggregated product data. Based on the user's specified criteria (price range, ratings, etc.), the process narrows down the results to only those products that match the criteria.

[0605] Step 10:

[0606] Server: Ranks the filtered products. Ranking criteria include price, popularity, review rating, etc. Generates a ranked product list.

[0607] Step 11:

[0608] Server: Sends the filtered and ranked product list to the device. Example data sent includes {"ranked_products": [...]}.

[0609] Step 12:

[0610] Device: The received product list is displayed to the user in a format appropriate for the user interface, such as list view or grid view.

[0611] Step 13:

[0612] Users can select the product they are interested in from the list of products presented and proceed to the detail page or purchase page. If they do not like the product, they can enter new search criteria and search again.

[0613] Step 14:

[0614] Server: Records user actions (selection, confirmation of details, purchase, etc.) and stores them as analytical data to improve search accuracy next time.

[0615] Step 15:

[0616] Server: Prepares to make personalized product suggestions based on the analyzed user data.

[0617] This series of steps allows users to easily find the products they are looking for and allows the system to continuously improve its accuracy.

[0618] Example 1

[0619] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0620] When searching for a wide variety of products on the Internet, modern consumers often spend a lot of time and effort trying to find the product they are looking for based on vague or abstract criteria. Furthermore, existing search engines and e-commerce platforms are unable to fully utilize users' past search history and purchasing behavior, resulting in a lack of personalized and appropriate product suggestions. This makes it difficult for users to efficiently find the products they need.

[0621] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0622] In this invention, the server includes means for receiving vague conditions entered by a user and transmitting them to the server, means for the server to analyze the received vague conditions based on a generative AI model and convert them into an appropriate search query, means for the server to crawl multiple e-commerce platforms and aggregate product data based on the converted search query, means for the server to filter the aggregated product data and rank it based on the user's specified conditions, means for the server to transmit the filtered and ranked product data to a terminal and for the terminal to display it to the user, and means for the terminal to present the received product data to the user in a display format such as a list view or grid view, thereby enabling users to efficiently find desired products even with vague conditions.

[0623] A "user" is a person who enters vague criteria and receives search results.

[0624] "Device" means a device used by a User to enter search criteria and view product data returned by the Server. Examples include smartphones, tablets, and PCs.

[0625] The "server" refers to a central computer that receives vague conditions sent by users, converts them into search queries using a generative AI model, and retrieves the necessary product data from multiple e-commerce platforms.

[0626] A "generative AI model" refers to an artificial intelligence technique that analyzes vague conditions and converts them into specific search queries.

[0627] A "search query" refers to a search condition that converts a user's vague conditions into a specific and optimal form.

[0628] An "e-commerce platform" refers to an online marketplace that offers goods and services over the web. Examples include Amazon, eBay, and Rakuten.

[0629] "Product Data" refers to information retrieved from e-commerce platforms based on a search query, including prices, descriptions, user reviews, and ratings.

[0630] "Filtering" refers to the process of narrowing down the acquired product data based on the user's specified conditions.

[0631] "Ranking" refers to the process of hierarchically arranging filtered product data based on criteria such as price, popularity, and review rating.

[0632] "List view" refers to a method of displaying product data in a vertical list format.

[0633] "Grid view" refers to a method of displaying product data in multiple rows and columns.

[0634] A "feedback loop" is a system that analyzes users' search results and purchasing behavior history and uses them to improve the accuracy of their next search.

[0635] In an embodiment of the present invention, a system is provided that allows users to easily search for desired products and obtain optimal results. This system is composed of a series of processes: acquiring user input, analyzing conditions, crawling e-commerce sites, filtering and ranking the results, presenting optimal search results, and providing personalized suggestions and a feedback loop.

[0636] Hardware and Software Configuration

[0637] User

[0638] Users enter vague criteria into the search bar on their devices (smartphones, tablets, PCs, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[0639] Terminal

[0640] The device receives the information entered by the user. It then converts the data into an appropriate data format, such as JSON, and sends it to the server. For example, the vague query entered by the user, "I want autographed merchandise by a famous athlete," can be converted to "{"query": "I want autographed merchandise by a famous athlete"}".

[0641] server

[0642] The server receives vague conditions sent from the device. This data is passed to a generative AI model (e.g., GPT-3). The generative AI model analyzes the user's vague conditions and converts them into specific search queries (e.g., "famous athlete signed merchandise"). The converted search queries are optimized for crawling e-commerce sites (e.g., Amazon, eBay, Rakuten).

[0643] The server crawls multiple e-commerce platforms using the search query generated by the generative AI model, sends the search query to each platform, and retrieves relevant product data, including prices, descriptions, user reviews, and ratings.

[0644] The server then aggregates the product data obtained from each e-commerce platform. This aggregated product data is first filtered and narrowed down based on user-specified criteria (e.g., price range, rating, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review rating.

[0645] The ranked product data is sent to the device, which then presents the received product data to the user in a display format such as list view or grid view, allowing users to easily find the best product from a large number of options.

[0646] Finally, the server analyzes the user's past purchasing behavior and search history to provide personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to refine the next search criteria. This allows the user to receive more personalized search results and product suggestions.

[0647] Specific operation example

[0648] For example, if a user wants to find autographed merchandise of a famous athlete and enters the search criteria "I want autographed merchandise of a famous athlete," the system will perform the following actions:

[0649] 1. Getting User Input

[0650] The user types "I want autographed merchandise of a famous athlete" into their device and clicks the "Search" button.

[0651] 2. Analysis of Conditions

[0652] The server passes the received input to a generative AI model, which generates a specific search query: "famous athlete signed merchandise."

[0653] 3. Crawling e-commerce sites

[0654] The server uses the generated search query to crawl multiple e-commerce platforms, such as Amazon, eBay, and Rakuten, to collect relevant product data.

[0655] 4. Filtering and Ranking Results

[0656] The server filters the collected product data based on price, user reviews, etc., and ranks them.

[0657] 5. Providing optimal search results

[0658] The server sends the filtered and ranked product data to the device, which then displays it to the user in a list view, showing each product's image, price, and rating.

[0659] 6. Personalized recommendations and feedback loops

[0660] The server analyzes the products the user clicks on and their purchase history, allowing it to suggest more suitable products for the next search.

[0661] Prompt Sentence Examples

[0662] As an example of a prompt statement, enter the following:

[0663] If a user types in "I want autographed merchandise of a famous athlete," analyze what kind of search query this translates into.

[0664] As described above, the present invention realizes a system that enables a user to efficiently find a desired product even if the user has vague conditions.

[0665] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0666] Step 1:

[0667] A user uses their device to enter vague search criteria into the search bar, such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[0668] Specifically, the user types text into an input field and performs a submit operation.

[0669] Input: Vague conditions entered by the user (e.g., "I want autographed merchandise by a famous athlete")

[0670] Output: Data when the send button is pressed and the search conditions are confirmed

[0671] Step 2:

[0672] After receiving the information entered by the user, the device converts the data into an appropriate data format, such as JSON, and sends it to the server. For example, it converts the data into JSON data such as "{'query': 'I want autographed merchandise of a famous athlete'}".

[0673] Specifically, the device converts the input text into a JSON string and sends it to the server as an HTTP request.

[0674] Input: User input data (e.g., "I want autographed merchandise from a famous athlete")

[0675] Output: Search criteria data in JSON format (e.g., "{'query': 'I want autographed merchandise by famous athletes'}")

[0676] Step 3:

[0677] The server receives the vague search terms sent from the device and passes this data to the generative AI model, which then analyzes the user's vague search terms and converts them into a specific search query (e.g., "famous athlete signed merchandise").

[0678] Specifically, the server calls an API endpoint and passes the data to the generative AI model for analysis.

[0679] Input: Search criteria data in JSON format (e.g., "{'query': 'I want autographed merchandise by famous athletes'}")

[0680] Output: A specific search query (e.g., "famous athlete signed merchandise")

[0681] Step 4:

[0682] The server crawls multiple e-commerce platforms using the search query generated by the generative AI model, sending the search query to each platform and retrieving relevant product data (e.g., price, description, user reviews, ratings, etc.).

[0683] Specifically, the server runs a crawling script and sends API requests to each platform to collect product data.

[0684] Input: A specific search query (e.g., "famous athlete signed merchandise")

[0685] Output: Product data obtained from each platform (e.g. price, description, user reviews, ratings)

[0686] Step 5:

[0687] The server aggregates product data obtained from each e-commerce platform. The aggregated product data is then filtered and narrowed down based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[0688] Specifically, the server executes code to sort and filter the product data.

[0689] Input: Raw data from each platform (e.g. price, description, user reviews, ratings)

[0690] Output: Filtered and ranked product data

[0691] Step 6:

[0692] The server transmits the filtered and ranked product data to the terminal, which presents the received product data to the user in a display format such as a list view or a grid view.

[0693] Specifically, the product list configured on the server side is sent to the terminal as an HTTP response, and the data is displayed as a UI on the terminal side.

[0694] Input: Filtered and ranked product data

[0695] Output: A list of products displayed in list view or grid view

[0696] Step 7:

[0697] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to improve the accuracy of the next search criteria. This provides the user with more personalized search results and product suggestions.

[0698] Specifically, the server accesses the user database and runs an algorithm that analyzes past data.

[0699] Input: User's past purchasing behavior and search history

[0700] Output: personalized product recommendations

[0701] (Application example 1)

[0702] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0703] Conventional e-commerce systems have the problem that it is difficult for users to find the desired product unless they enter precise keywords. Furthermore, searches using vague conditions often result in the display of many irrelevant products, making them inefficient for users. Meanwhile, individual suggestions based on purchase and search history are not fully utilized, making it difficult to suggest optimal products tailored to the user's needs. Furthermore, the limited options for displayed products and the lack of feedback reflected in the next search also detract from the user experience.

[0704] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0705] In this invention, the server includes a means for receiving vague search terms entered by a user, analyzing them based on a generative AI model, and converting them into an appropriate search query; a means for crawling multiple e-commerce platforms and aggregating product data based on the converted search query; and a means for filtering the aggregated product data and ranking them based on the user's specified search terms. This allows users to search for optimal products even with vague search terms. The server also uses the generative AI model to analyze past purchasing behavior and make personalized product suggestions, thereby providing more appropriate products to users. Additionally, the system includes a means for the terminal to allow users to select additional products based on the displayed search results, and a means for sending user feedback to the server to create a feedback loop to improve the accuracy of the next search, thereby improving the user experience.

[0706] "User input" refers to the action of a user providing their desired conditions or requests through an electronic device.

[0707] A "generative AI model" is an algorithm or system that uses artificial intelligence technology to convert vague conditions into specific search queries.

[0708] A "search query" is a keyword or phrase that a user enters into a search engine to retrieve specific information.

[0709] "E-commerce platform" refers to a website or application for buying and selling products and services over the Internet.

[0710] "Crawling" is the process of automatically visiting web pages on the Internet and collecting data.

[0711] "Product data" is a data set that contains detailed product information, such as price, description, and reviews.

[0712] "Filtering" is a process for narrowing down data based on specific conditions.

[0713] "Ranking" is the process of ranking data based on specific criteria.

[0714] "Personalized suggestions" analyze a user's past behavior and preferences to suggest products and services that are best suited to each individual user.

[0715] A "feedback loop" is a system that collects user reactions and behavior and reflects them in the next service provision.

[0716] The system for realizing this invention mainly includes a terminal that processes user input, a server that analyzes ambiguous conditions and generates optimal search queries, and a crawling function that aggregates product data from e-commerce platforms based on the generated queries.

[0717] User input via terminal

[0718] The user inputs their desired conditions and requests into the terminal. For example, the user might input "I want autographed merchandise of a famous athlete." Once the input is confirmed, the terminal sends this information to the server. The terminal can be a smartphone, tablet, PC, or other device.

[0719] Server-based analysis of ambiguous conditions

[0720] The server passes the vague conditions received from the device to a generative AI model. For example, the BART transformer model is used as the generative AI model. The model converts the vague conditions into a specific search query. For example, the condition "I want signed merchandise by a famous athlete" is converted to "famous athlete signed merchandise."

[0721] Crawling product data from e-commerce sites

[0722] The server uses the generated search query to crawl multiple e-commerce platforms (e.g., Amazon and Rakuten) to retrieve relevant product data, including information such as price, description, reviews, and ratings.

[0723] Filtering and Ranking Data

[0724] The acquired product data is filtered on the server and ranked based on user-specified criteria (price range, ratings, etc.) The filtering and ranking algorithms are implemented in Python, for example.

[0725] Providing optimal search results

[0726] The server sends the filtered and ranked product data to the device, which displays it to the user, who can then browse the best products in a list or grid view and select specific products.

[0727] Personalized product recommendations

[0728] The server analyzes the user's past purchasing behavior and search history and uses a generative AI model to make personalized suggestions, thereby providing users with more relevant and personalized product suggestions.

[0729] Formation of a feedback loop

[0730] The device sends user operation information and feedback to the server, which then analyzes the feedback and creates a feedback loop to improve the accuracy of the next search.

[0731] Specific examples

[0732] For example, if a user types in "latest smartphone case recommendations," the server converts this to "latest smartphone cases recommendations," retrieves smartphone case information from Amazon and Rakuten, and displays a list.

[0733] Prompt Sentence Examples

[0734] "Generate what query would be converted when a user types 'I want autographed merchandise of a famous athlete' into a product search app."

[0735] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0736] Step 1:

[0737] Get the conditions entered by the user

[0738] The device accepts input from the user. Specifically, the user enters vague keywords such as "I want autographed merchandise of a famous athlete" into the product search bar. Once this input is confirmed, the device converts the information into a data format such as JSON and sends the data to the server.

[0739] Input: Vague keywords entered by the user

[0740] Output: Keyword information converted into data format such as JSON

[0741] Step 2:

[0742] Convert vague conditions into specific search queries

[0743] The server passes the vague conditions received from the device to a generative AI model. The generative AI model (e.g., the BART transformer model) analyzes the conditions and converts them into specific search queries. For example, "I want autographed merchandise by a famous athlete" is converted to "famous athlete signed merchandise."

[0744] Input: Vague keywords in JSON format sent from the terminal

[0745] Output: A specific search query transformed by the generative AI model

[0746] Step 3:

[0747] Crawl multiple e-commerce platforms based on a search query

[0748] The server uses the generated search query to access the APIs of multiple e-commerce platforms (e.g., Amazon and Rakuten) and crawl product data, including product names, prices, descriptions, reviews, ratings, etc.

[0749] Input: A specific search query transformed by a generative AI model

[0750] Output: Product data obtained from multiple e-commerce platforms

[0751] Step 4:

[0752] Filter and rank your product data

[0753] The server aggregates the acquired product data and filters it based on the user's specified conditions (price range, ratings, etc.), and then ranks the filtered product data based on criteria such as ratings and price.

[0754] Input: Product data obtained from multiple e-commerce platforms

[0755] Output: Filtered and ranked product data

[0756] Step 5:

[0757] Displaying the most relevant product information to users

[0758] The server transmits the filtered and ranked product data to the terminal, which displays the data to the user in a format such as a list view or a grid view, allowing the user to review and select products.

[0759] Input: Filtered and ranked product data

[0760] Output: Product information displayed in list or grid view format

[0761] Step 6:

[0762] Make personalized product recommendations based on past purchasing behavior and search history

[0763] The server analyzes the user's past purchasing behavior and search history, and uses a generative AI model based on that information to make personalized product suggestions, allowing users to be recommended products that match their preferences.

[0764] Input: User's past purchasing behavior and search history

[0765] Output: personalized product recommendations

[0766] Step 7:

[0767] Forming a feedback loop

[0768] The device sends user operation information and feedback to the server, which analyzes this feedback and reflects it in order to improve the accuracy of the next search, allowing the user to obtain more accurate search results.

[0769] Input: User operation information and feedback

[0770] Output: Data reflected to improve search accuracy next time

[0771] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0772] In an embodiment of the present invention, a system is provided that allows users to easily search for desired products using vague search criteria and obtain optimal results. Furthermore, the system recognizes the user's emotions and improves the accuracy of search queries based on those emotions, thereby providing more personalized product offers. A specific embodiment of the system and its operation are described below.

[0773] Getting User Input

[0774] User

[0775] Users enter vague criteria into the search bar on their device (smartphone, tablet, PC, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." Once they've finished entering the information, they click the "Search" button.

[0776] Terminal

[0777] The terminal receives the information entered by the user, converts the data into an appropriate data format such as JSON, and sends it to the server.

[0778] Condition analysis and emotion recognition

[0779] server

[0780] The server receives the ambiguous condition data sent from the device, passes this data to the generative AI model, and begins the analysis process.

[0781] Generative AI Models

[0782] The generative AI model analyzes the user's vague search criteria and converts them into a specific search query, which is then used to provide optimal search results.

[0783] Emotion Engine

[0784] The emotion engine analyzes the user's emotions at the time of input from voice, text input, facial expressions, etc. Based on the results of this analysis, the generative AI model optimizes its condition analysis. For example, if the user is excited, it will make more proactive product suggestions, while if the user is calm, it will emphasize more detailed information.

[0785] Crawling e-commerce sites

[0786] server

[0787] The server uses the generated search query and the sentiment data obtained from the sentiment engine to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.) and sends a search query to each platform to retrieve related product data.

[0788] Filtering and ranking results

[0789] server

[0790] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. This dataset is then filtered to narrow down the list based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[0791] Providing optimal search results

[0792] server

[0793] The server transmits the filtered and ranked product data to the terminal.

[0794] Terminal

[0795] The device displays the received product data to the user in a list view, grid view, or other display format, allowing the user to easily find the best product from a wide range of options.

[0796] Personalized recommendations and feedback loops

[0797] server

[0798] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is stored and used to improve the accuracy of the next search criteria. User emotional data is also stored and reflected in future searches and suggestions.

[0799] user

[0800] Users can select a product from the product list displayed on their device and proceed to the details page or purchase page. They can also enter new search criteria and search again.

[0801] Specific examples

[0802] For example, suppose a user who wants to find autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete" and seems to be having fun. In this case, the system will perform the following actions.

[0803] 1. User Input

[0804] The user types "I want autographed merchandise from a famous athlete" into the device.

[0805] 2. Condition Analysis and Emotion Recognition

[0806] The server passes the input to a generative AI model for analysis, resulting in the search query "famous athlete signed merchandise."

[0807] The emotion engine analyzes the user's positive emotions and, based on the generated query, prioritizes products with more positive reviews.

[0808] 3. E-commerce site crawling

[0809] The server crawls multiple e-commerce platforms and collects product data.

[0810] 4. Filtering and Ranking Results

[0811] The server filters the collected product data and prioritizes and ranks products that have high ratings and many positive reviews that reflect the user's enjoyment emotions.

[0812] 5. Search result presentation

[0813] The server sends the filtered and ranked results to the device.

[0814] The device displays the appropriate product list to the user.

[0815] 6. Personalized Offers

[0816] The server analyzes the user's behavioral and emotional data to make more relevant personalized suggestions in the future.

[0817] In this way, a system is provided that allows users to easily find the products they are looking for without knowing the specific product names or detailed conditions, and furthermore, allows them to receive appropriate product suggestions based on their emotions.

[0818] The processing flow will be explained below.

[0819] Step 1:

[0820] User: Enters vague criteria such as "I want autographed merchandise of a famous athlete" into the device's search bar. Once entered, presses the "Search" button.

[0821] Step 2:

[0822] Terminal: Receives the search criteria entered by the user, converts the data into an appropriate format (e.g., JSON), and sends it to the server. An example of sending data is {"query":"I want autographed merchandise by a famous athlete"}.

[0823] Step 3:

[0824] Server: Receives search criteria data sent from the device and temporarily stores this data in memory.

[0825] Step 4:

[0826] Server: Passes the received data to the generative AI model and begins the analysis process. The generative AI model analyzes the user's vague search criteria and converts them into an appropriate search query (e.g., "famous athlete signed merchandise").

[0827] Step 5:

[0828] Emotion engine: Analyzes the emotions at the time of input from the user's voice, text input, facial expressions, etc. The emotion engine outputs the analysis results and optimizes the condition analysis of the generative AI model.

[0829] Step 6:

[0830] Server: Stores the search queries generated by the generative AI model and the sentiment analysis data from the sentiment engine, and prepares them to be sent to multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.).

[0831] Step 7:

[0832] Server: Uses generated search queries to access APIs and crawl the web for each e-commerce platform, and collects product data (price, description, ratings, etc.) from each platform.

[0833] Step 8:

[0834] Each e-commerce site: Returns a list of products based on the search query to the server through the provided API or web interface. For example, the results are returned in the format {"site": "Amazon", "results": [...]}.

[0835] Step 9:

[0836] Server: Aggregates product data collected from each e-commerce platform to form a large product dataset, which is temporarily stored in memory.

[0837] Step 10:

[0838] Server: Filters the aggregated product data. Based on the user's criteria (price range, ratings, etc.), it runs a process to narrow down the list to only products that match the criteria. It also prioritizes products based on sentiment data (e.g., if the user seems happy, prioritize products with many positive reviews).

[0839] Step 11:

[0840] Server: Ranks the filtered products. Ranking criteria include price, popularity, review rating, etc. Generates a ranked product list.

[0841] Step 12:

[0842] Server: Sends the filtered and ranked product list to the device. Example data sent includes {"ranked_products": [...]}.

[0843] Step 13:

[0844] Device: The received product list is displayed to the user in a format appropriate for the user interface, such as list view or grid view.

[0845] Step 14:

[0846] Users can select the product they are interested in from the list of products presented and proceed to the detail page or purchase page. If they do not like the product, they can enter new search criteria and search again.

[0847] Step 15:

[0848] Server: Records user behavior (selection, confirmation of details, purchase, etc.) and emotional data, and stores it as analytical data to improve search accuracy next time.

[0849] Step 16:

[0850] Server: Prepares to make future personalized product recommendations based on the analyzed user and sentiment data.

[0851] This series of steps allows users to easily find the products they are looking for and allows the system to continuously improve its accuracy.

[0852] Example 2

[0853] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0854] In conventional e-commerce platforms, when users search for products using vague criteria, they often encounter problems such as insufficient search results or being unable to find the desired product. Furthermore, because the platform does not take into account the user's purchasing motivation or emotions, personalized product suggestions are not possible, often resulting in low satisfaction. The present invention aims to solve these problems and provide users with optimal product search results and personalized suggestions.

[0855] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0856] In this invention, the server includes a means for analyzing vague search terms entered by a user based on a generative AI model and converting them into an appropriate search query, an emotion recognition means for analyzing the user's emotional state, and a means for crawling multiple e-commerce platforms and aggregating product data based on the converted search query. This allows a specific search query appropriate to the user's intent to be generated, even if the user enters vague search terms, and the aggregated product data can be ranked based on the user's specified search terms and emotional state. Furthermore, the system analyzes the user's past behavioral data and emotional data to provide personalized product suggestions, thereby increasing user satisfaction.

[0857] "Ambiguous search conditions entered by the user" refers to non-specific search conditions entered by the user through the terminal.

[0858] A "generative AI model" refers to an artificial intelligence model that analyzes vague conditions entered by users and converts them into appropriate search queries.

[0859] A "search query" refers to a specific search term or phrase that a generative AI model generates by analyzing ambiguous conditions.

[0860] "Emotion recognition means" refers to technology that analyzes the emotional state of a user from their input or interaction, including, for example, voice, text input, and facial expression recognition.

[0861] An "e-commerce platform" is a website or application that enables the buying and selling of goods and services online, such as Amazon or eBay.

[0862] "Crawling" refers to the act of a server automatically searching for information on a particular website and collecting data.

[0863] "Product Data" refers to all product-related information obtained from the e-commerce platform, such as product name, price, review rating, etc.

[0864] "Filtering" refers to the process of narrowing down collected product data based on user-specified conditions.

[0865] "Ranking" refers to the process of ranking filtered product data based on specific criteria, such as price, popularity, or review rating.

[0866] "Specified conditions" refer to conditions that a user sets to narrow down search results, such as a price range or rating range.

[0867] "Individualized product proposals" refer to product proposals that are optimized for each individual user based on the user's past behavioral and emotional data.

[0868] A "feedback loop" refers to a system that receives feedback from users and uses that information to improve the accuracy of subsequent searches and suggestions.

[0869] The present invention is a system for providing optimal search results when a user searches for products under vague conditions. Furthermore, the system recognizes the user's emotions and improves the accuracy of the search query based on the emotions, thereby providing more personalized product suggestions. A specific embodiment of the system and its operation are described below.

[0870] 1. Getting User Input

[0871] User

[0872] Users enter vague criteria into the search bar on their device (smartphone, tablet, PC, etc.), for example, an abstract request such as "I want autographed merchandise of a famous athlete." Once they're done, they click the "Search" button.

[0873] Terminal

[0874] The device acquires the conditions entered by the user, converts this input data into an appropriate data format such as JSON, and sends it to the server. At this time, when the user has completed entering the information on the device, the device sends an HTTP request to the server.

[0875] 2. Condition analysis and emotion recognition

[0876] server

[0877] The server receives vague condition data sent from the device. This data is passed to the generative AI model as a prompt. For example, the data {"query": "I want autographed merchandise of a famous athlete"} is sent.

[0878] Generative AI Models

[0879] The generative AI model analyzes the received prompt and converts the user's vague search criteria into a specific search query, such as "famous athlete signed merchandise." This converted query is cached on the server and can be reused.

[0880] Emotion Engine

[0881] The emotion engine analyzes user input (voice, text, facial expressions, etc.) to recognize the user's emotional state. For example, facial recognition technology can be used to determine whether a user is happy when typing. This emotional data influences the analysis of the generative AI model.

[0882] 3. Crawling e-commerce sites

[0883] server

[0884] The server uses the converted search query and sentiment data to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.). For each platform, it sends a new search query and collects related product data. For example, it uses the Amazon API to send the query "famous athlete signed merchandise" and retrieves the corresponding product data.

[0885] 4. Filtering and Ranking Results

[0886] server

[0887] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. This dataset is then filtered to narrow down the results based on user-specified criteria (price range, ratings, etc.). Specifically, it selects highly rated products within a certain price range. It also prioritizes products with proactive product suggestions and detailed information based on emotional data. For example, for users with a positive emotional state, it prioritizes products with many positive reviews.

[0888] 5. Providing optimal search results

[0889] server

[0890] The server sends the filtered and ranked product data to the terminal, for example, returning a product list in JSON format.

[0891] Terminal

[0892] The device displays the received product data to the user. The display format can be selected from list view, grid view, etc. For example, in list view, the product name, image, price, and review rating are displayed.

[0893] 6. Personalized recommendations and feedback loops

[0894] server

[0895] The server analyzes users' past purchasing behavior and search history to make personalized product suggestions. It also accumulates behavioral data, such as when users click on a specific product to view its details or make a purchase. This data is analyzed and saved along with emotional data to help improve the accuracy of the next search.

[0896] User

[0897] Users can select products they are interested in from the products displayed on their devices, proceed to the details and purchase process, or enter new search criteria and search again.

[0898] Specific examples

[0899] For example, suppose a user searching for autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete" and seems to be having fun. In this case, the system performs the following specific actions:

[0900] 1. User Input

[0901] The user types "I want autographed merchandise from a famous athlete" into the device.

[0902] The user clicks the Search button.

[0903] 2. Condition Analysis and Emotion Recognition

[0904] The server passes the input data to the generative AI model.

[0905] A generative AI model generates the search query "famous athlete signed merchandise."

[0906] The emotion engine analyzes the user's happy facial expression.

[0907] 3. E-commerce site crawling

[0908] A server sends search queries to multiple e-commerce platforms to collect product data.

[0909] 4. Filtering and Ranking Results

[0910] The server filters the collected product data and prioritizes products with positive reviews and high ratings.

[0911] 5. Search result presentation

[0912] The server transmits the ranked product data to the terminal.

[0913] The device displays the product list to the user in a list view format.

[0914] 6. Personalized Offers

[0915] The server analyzes the user's behavioral and emotional data to make more appropriate personalized suggestions in the future.

[0916] Example prompt sentence:

[0917] "I want autographed merchandise from famous athletes."

[0918] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0919] Step 1:

[0920] Getting User Input

[0921] The user enters vague search criteria into the search bar of their device (smartphone, tablet, PC, etc.). For example, they might type "I want autographed merchandise of a famous athlete." Then, they click the "Search" button. The input data is in text format.

[0922] Step 2:

[0923] Sending data from the device to the server

[0924] The device converts the user-entered criteria into an appropriate data format, such as JSON. The device then sends this data to the server using an HTTP request, such as {"query": "I want autographed merchandise by a famous athlete"}. This data is provided by the user as input.

[0925] Step 3:

[0926] Condition Analysis

[0927] The server passes the vague condition data received from the device to the generative AI model. The generative AI model analyzes this data and converts it into a specific search query. For example, the vague condition "I want signed merchandise by a famous athlete" is converted into the specific search query "famous athlete signed merchandise." This converted query is output as search data.

[0928] Step 4:

[0929] Emotion Analysis

[0930] The server uses an emotion engine to analyze the user's input format (voice, text, facial expression, etc.). The emotion engine analyzes emotions from the user's facial expression and voice data at the time of input and returns the results to the server. For example, facial expression analysis can yield an emotion result of "looking happy." This emotion data becomes input for use in optimizing search queries.

[0931] Step 5:

[0932] Crawling e-commerce sites

[0933] The server uses the generated search query and the analyzed sentiment data to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.). During the crawling, it uses the platform's API or scraping technology to send the search query and retrieve the corresponding product data. For example, it sends the search query "famous athlete signed merchandise" to Amazon to retrieve a list of related products. This product data is output as a result from the e-commerce site.

[0934] Step 6:

[0935] Filtering and ranking results

[0936] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. Products are filtered from this dataset based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on emotion data. For example, products with many positive reviews are prioritized for users with positive emotions. The filtered and ranked product data is output as an optimal product list.

[0937] Step 7:

[0938] Providing optimal search results

[0939] The server sends the filtered and ranked product data to the device, for example in the format {"products": [listed product data]}. This data is returned to the user as the best search results.

[0940] Terminal

[0941] The device displays the received product data to the user. The display format can be selected from list view, grid view, etc. For example, list view displays the product name, image, price, and review rating. This displayed product information is provided in a format that is easy for the user to view.

[0942] Step 8:

[0943] Personalized recommendations and feedback loops

[0944] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. It also accumulates behavioral data, such as when a user clicks on a specific product to view its details page or make a purchase. This data is analyzed to help improve the accuracy of the next search criteria. For example, if a user repeatedly takes positive action regarding "signature goods," similar products will be included in the next suggestion. This feedback data is output as the user's behavioral pattern.

[0945] User

[0946] Users can select products they are interested in from the displayed products and proceed to the purchase process. They can also search again by entering new search criteria, improving the user experience. These new search criteria are then processed as data to be input again into the system.

[0947] As described above, the system of the present invention can efficiently carry out a series of processes from vague conditions to specific product proposals.

[0948] (Application example 2)

[0949] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0950] Existing search systems have difficulty finding appropriate search results for vague user inputs, making it difficult for users to smoothly find the products they are looking for. Furthermore, the lack of personalized suggestions based on user emotions and past behavior makes it difficult to provide a more personalized shopping experience. To solve these problems, there is a need for a search system that takes user emotions into account.

[0951] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving ambiguous conditions entered by a user and transmitting them to the server, means for the server to analyze the ambiguous conditions received based on a generative AI model and convert them into an appropriate search query, means for the server to crawl multiple e-commerce platforms and aggregate product data based on the converted search query, means for the server to filter the aggregated product data using an emotion recognition engine taking into account the user's emotional data and rank it based on the user's specified conditions, and means for the server to transmit the filtered and ranked product data to a terminal, which then displays it to the user. This enables users to accurately search for products even with ambiguous conditions and receive more personalized product suggestions based on their emotions.

[0952] "User" means an individual or organization that uses the System to search for and purchase products.

[0953] "Vague terms" are abstract search queries or requests entered by users without specifying specific products or specifications.

[0954] "Server" refers to a computer system that receives input data from users and performs processes such as analysis using generative AI models, emotion recognition engines, crawling e-commerce platforms, filtering and ranking data, etc.

[0955] A "generative AI model" is a type of artificial intelligence that analyzes ambiguous conditions and converts them into appropriate search queries, and is based on natural language processing.

[0956] An "emotion recognition engine" is a technology for analyzing emotions from non-verbal information such as a user's facial expressions and voice.

[0957] An "e-commerce platform" is an online marketplace for buying and selling goods and services over the Internet.

[0958] "Crawling" is the process of automatically gathering information on a website.

[0959] "Filtering" is the process of removing unnecessary data from collected product data based on specific criteria and extracting only the necessary data.

[0960] "Ranking" is the process of ranking the filtered product data based on specific evaluation criteria.

[0961] A "terminal" is a device that a user uses to access the system, including smartphones and computers.

[0962] A "feedback loop" is a cyclical process that analyzes information collected from users (e.g., search history, sentiment data) to improve the accuracy of future search criteria and suggestions.

[0963] The system for implementing this invention allows users to search for products under vague conditions and optimizes the results based on emotions. The specific configuration and operation of this system will be described below.

[0964] System Configuration

[0965] Hardware

[0966] Smartphone: A device with a built-in camera and microphone that allows users to access the system.

[0967] Server: A computer system equipped with a high-speed processor and capable of processing large amounts of data. Cloud-based operation is recommended.

[0968] software

[0969] Generative AI model: A natural language processing model to convert users' vague search terms into specific search queries. In this case, we use OpenAI's GPT-4.

[0970] Emotion recognition engine: An engine that analyzes the user's facial expressions and voice to recognize emotions. It uses Google Cloud Vision API and Microsoft Azure Face API.

[0971] E-commerce site crawling library: A library for collecting product data from e-commerce platforms. It uses BeautifulSoup (Python) and other tools.

[0972] Cloud platform: Infrastructure for operations and data processing. We recommend using Amazon Web Services (AWS).

[0973] System Operation

[0974] 1. Getting User Input

[0975] Users enter vague terms into the search bar on their smartphone and press the search button.

[0976] Example: "I want new shoes."

[0977] 2. Emotion recognition

[0978] The device's camera and microphone are used to collect the user's facial expressions and voice.

[0979] The emotion recognition engine analyzes this data and recognizes the user's emotions.

[0980] Example: A user looks happy.

[0981] 3. Condition analysis

[0982] Vague search queries are fed into a generative AI model, which converts them into specific search queries.

[0983] Example: "new shoes"

[0984] 4. Crawling e-commerce sites

[0985] The server uses the generated search query to crawl multiple e-commerce platforms to collect product data.

[0986] The collected data is converted into a unified format and aggregated on the server.

[0987] 5. Filtering and Ranking Results

[0988] The aggregated product data is filtered taking into account user sentiment data.

[0989] The filtered data is then ranked based on criteria such as price, popularity, and review ratings.

[0990] 6. Presentation of search results

[0991] The server transmits the filtered and ranked data to the terminal.

[0992] The terminal displays a list of products that are most suitable for the user.

[0993] 7. Creating feedback loops

[0994] User selections, search history, emotional data, etc. are stored on the server to improve search accuracy from the next time onwards.

[0995] Prompt Sentence Examples

[0996] Here is an example of a prompt for the generative AI model:

[0997] A user searched for: "I want new shoes." The user's sentiment seems to be happy. Convert this into a specific search query.

[0998] In this way, users can easily find the products they are looking for without having to enter specific product names or detailed search criteria. Furthermore, the system's emotion-based product suggestions enable a more personalized shopping experience.

[0999] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1000] Step 1: Getting User Input

[1001] The user enters vague search criteria into the device, for example, "I want new shoes." Once the search is complete, the user presses the search button. The device converts the input data into JSON format and sends it to the server.

[1002] Input: User's vague search criteria

[1003] Output: Input data in JSON format

[1004] Specific actions: Enter search criteria in the search bar and press the button.

[1005] Step 2: Emotion Recognition

[1006] The device uses a built-in camera and microphone to collect the user's facial expressions and voice. This data is sent to an emotion recognition engine to analyze the user's emotions. The analysis results are sent to the server in JSON format.

[1007] Input: User's facial expression and voice data

[1008] Output: Emotion data in JSON format

[1009] Specific actions: Capture facial expressions and voice using a camera and microphone.

[1010] Step 3: Condition analysis

[1011] The server receives the search criteria and emotion data sent from the device, passes this data to the generative AI model, and converts it into a specific search query. The generative AI model analyzes the prompt and generates an appropriate search query.

[1012] Input: Fuzzy search terms and sentiment data in JSON format

[1013] Output: Specific search query

[1014] Specific operation: The generative AI model analyzes the prompt sentence and generates a search query.

[1015] Step 4: Crawl your e-commerce site

[1016] The server crawls multiple e-commerce platforms using the generated specific search query, collects product data using libraries such as BeautifulSoup, converts the collected data into a unified format, and aggregates it on the server.

[1017] Input: A specific search query

[1018] Output: Aggregated product data in a unified format

[1019] Specific operation: Collects product data using e-commerce site crawl libraries.

[1020] Step 5: Filter and rank the results

[1021] The server receives the aggregated product data and filters it using an emotion recognition engine, taking into account the user's emotional data. The filtered data is then ranked based on criteria such as price, ratings, and reviews.

[1022] Input: Aggregated product data and sentiment data in a unified format

[1023] Output: Filtered and ranked product data

[1024] Specific behavior: Analyze product data and optimize based on user sentiment.

[1025] Step 6: Presenting search results

[1026] The server sends the filtered and ranked product data to the device, which then visually displays this data to the user in a format such as a list view or grid view.

[1027] Input: Filtered and ranked product data

[1028] Output: The optimal product list shown to the user

[1029] Specific operation: Display the product list on the terminal screen.

[1030] Step 7: Creating a feedback loop

[1031] The server stores user behavioral history and emotional data to improve the accuracy of search criteria and suggestions from the next time onward, allowing users to receive more personalized search results.

[1032] Input: User behavior history, emotional data

[1033] Output: Feedback to improve accuracy using accumulated data

[1034] Specific actions: Accumulate behavioral history and emotional data and reflect them in the next proposal.

[1035] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1036] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1037] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1038] [Third embodiment]

[1039] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1040] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1041] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1042] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1043] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1044] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1045] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1046] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1047] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1048] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1049] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1050] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[1051] In the embodiment of the present invention, a system is provided that allows users to easily search for products they want and obtain optimal results. A specific embodiment of this system and its operation will be described below.

[1052] Getting User Input

[1053] User

[1054] Users enter vague criteria into the search bar on their devices (smartphones, tablets, PCs, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[1055] Terminal

[1056] The terminal accepts information entered by the user, converts the data into an appropriate data format such as JSON, and sends it to the server.

[1057] Condition Analysis

[1058] server

[1059] The server receives vague conditions sent from the device. This data is passed to the generative AI model. The generative AI model analyzes the user's vague conditions and converts them into a specific search query. This converted search query is optimized to search for the product the user is looking for.

[1060] Crawling e-commerce sites

[1061] server

[1062] The server crawls multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.) using the search query generated by the generative AI model. It sends the search query to each platform and retrieves relevant product data, including prices, descriptions, user reviews, and ratings.

[1063] Filtering and ranking results

[1064] server

[1065] The server aggregates product data obtained from each e-commerce platform. The aggregated product data is filtered and narrowed down based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[1066] Providing optimal search results

[1067] server

[1068] The server transmits the filtered and ranked product data to the terminal.

[1069] Terminal

[1070] The device displays the received product data to the user in a list view, grid view, or other display format, allowing the user to easily find the best product from a wide range of options.

[1071] Personalized recommendations and feedback loops

[1072] server

[1073] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to improve the accuracy of the next search criteria. This provides the user with more personalized search results and product suggestions.

[1074] user

[1075] Users can select a product from the product list displayed on their device and proceed to the details page or purchase page. They can also enter new search criteria and search again.

[1076] Specific examples

[1077] For example, if a user who wants to find autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete," the system will perform the following actions.

[1078] 1. User Input

[1079] The user types "I want autographed merchandise from a famous athlete" into the device.

[1080] 2. Condition analysis

[1081] The server passes the input to a generative AI model for analysis, resulting in the search query "famous athlete signed merchandise."

[1082] 3. E-commerce site crawling

[1083] The server crawls multiple e-commerce platforms and collects product data.

[1084] 4. Filtering and Ranking Results

[1085] The server filters the collected product data and ranks it based on criteria such as price and ratings.

[1086] 5. Search result presentation

[1087] The server sends the filtered and ranked results to the device.

[1088] The device displays the appropriate product list to the user.

[1089] 6. Personalized Offers

[1090] The server analyzes user behavior and makes more relevant personalized offers in the future.

[1091] In this way, a system is provided that allows users to easily find the product they are looking for, even if they do not know the specific product name or detailed conditions.

[1092] The processing flow will be explained below.

[1093] Step 1:

[1094] User: Enters vague criteria such as "I want autographed merchandise of a famous athlete" into the device's search bar. Once entered, presses the "Search" button.

[1095] Step 2:

[1096] Terminal: Receives the search criteria entered by the user, converts the data into an appropriate format (e.g., JSON), and sends it to the server. An example of sending data is {"query":"I want autographed merchandise by a famous athlete"}.

[1097] Step 3:

[1098] Server: Receives search criteria data sent from the device and temporarily stores this data in memory.

[1099] Step 4:

[1100] Server: Passes the received data to the generative AI model and begins the analysis process. The generative AI model analyzes the user's vague search criteria and converts them into an appropriate search query (e.g., "famous athlete signed merchandise").

[1101] Step 5:

[1102] Server: Stores the search queries generated by the generative AI model and prepares them to be sent to multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten).

[1103] Step 6:

[1104] Server: Uses generated search queries to access APIs and crawl the web for each e-commerce platform, and collects product data (price, description, ratings, etc.) from each platform.

[1105] Step 7:

[1106] Each e-commerce site: Returns a list of products based on the search query to the server through the provided API or web interface. For example, the results are returned in the format {"site": "Amazon", "results": [...]}.

[1107] Step 8:

[1108] Server: Aggregates product data collected from each e-commerce platform to form a large product dataset, which is temporarily stored in memory.

[1109] Step 9:

[1110] Server: Filters the aggregated product data. Based on the user's specified criteria (price range, ratings, etc.), the process narrows down the results to only those products that match the criteria.

[1111] Step 10:

[1112] Server: Ranks the filtered products. Ranking criteria include price, popularity, review rating, etc. Generates a ranked product list.

[1113] Step 11:

[1114] Server: Sends the filtered and ranked product list to the device. Example data sent includes {"ranked_products": [...]}.

[1115] Step 12:

[1116] Device: The received product list is displayed to the user in a format appropriate for the user interface, such as list view or grid view.

[1117] Step 13:

[1118] Users can select the product they are interested in from the list of products presented and proceed to the detail page or purchase page. If they do not like the product, they can enter new search criteria and search again.

[1119] Step 14:

[1120] Server: Records user actions (selection, confirmation of details, purchase, etc.) and stores them as analytical data to improve search accuracy next time.

[1121] Step 15:

[1122] Server: Prepares to make personalized product suggestions based on the analyzed user data.

[1123] This series of steps allows users to easily find the products they are looking for and allows the system to continuously improve its accuracy.

[1124] Example 1

[1125] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1126] When searching for a wide variety of products on the Internet, modern consumers often spend a lot of time and effort trying to find the product they are looking for based on vague or abstract criteria. Furthermore, existing search engines and e-commerce platforms are unable to fully utilize users' past search history and purchasing behavior, resulting in a lack of personalized and appropriate product suggestions. This makes it difficult for users to efficiently find the products they need.

[1127] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1128] In this invention, the server includes means for receiving vague conditions entered by a user and transmitting them to the server, means for the server to analyze the received vague conditions based on a generative AI model and convert them into an appropriate search query, means for the server to crawl multiple e-commerce platforms and aggregate product data based on the converted search query, means for the server to filter the aggregated product data and rank it based on the user's specified conditions, means for the server to transmit the filtered and ranked product data to a terminal and for the terminal to display it to the user, and means for the terminal to present the received product data to the user in a display format such as a list view or grid view, thereby enabling users to efficiently find desired products even with vague conditions.

[1129] A "user" is a person who enters vague criteria and receives search results.

[1130] "Device" means a device used by a User to enter search criteria and view product data returned by the Server. Examples include smartphones, tablets, and PCs.

[1131] The "server" refers to a central computer that receives vague conditions sent by users, converts them into search queries using a generative AI model, and retrieves the necessary product data from multiple e-commerce platforms.

[1132] A "generative AI model" refers to an artificial intelligence technique that analyzes vague conditions and converts them into specific search queries.

[1133] A "search query" refers to a search condition that converts a user's vague conditions into a specific and optimal form.

[1134] An "e-commerce platform" refers to an online marketplace that offers goods and services over the web. Examples include Amazon, eBay, and Rakuten.

[1135] "Product Data" refers to information retrieved from e-commerce platforms based on a search query, including prices, descriptions, user reviews, and ratings.

[1136] "Filtering" refers to the process of narrowing down the acquired product data based on the user's specified conditions.

[1137] "Ranking" refers to the process of hierarchically arranging filtered product data based on criteria such as price, popularity, and review rating.

[1138] "List view" refers to a method of displaying product data in a vertical list format.

[1139] "Grid view" refers to a method of displaying product data in multiple rows and columns.

[1140] A "feedback loop" is a system that analyzes users' search results and purchasing behavior history and uses them to improve the accuracy of their next search.

[1141] In an embodiment of the present invention, a system is provided that allows users to easily search for desired products and obtain optimal results. This system is composed of a series of processes: acquiring user input, analyzing conditions, crawling e-commerce sites, filtering and ranking the results, presenting optimal search results, and providing personalized suggestions and a feedback loop.

[1142] Hardware and Software Configuration

[1143] User

[1144] Users enter vague criteria into the search bar on their devices (smartphones, tablets, PCs, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[1145] Terminal

[1146] The device receives the information entered by the user. It then converts the data into an appropriate data format, such as JSON, and sends it to the server. For example, the vague query entered by the user, "I want autographed merchandise by a famous athlete," can be converted to "{"query": "I want autographed merchandise by a famous athlete"}".

[1147] server

[1148] The server receives vague conditions sent from the device. This data is passed to a generative AI model (e.g., GPT-3). The generative AI model analyzes the user's vague conditions and converts them into specific search queries (e.g., "famous athlete signed merchandise"). The converted search queries are optimized for crawling e-commerce sites (e.g., Amazon, eBay, Rakuten).

[1149] The server crawls multiple e-commerce platforms using the search query generated by the generative AI model, sends the search query to each platform, and retrieves relevant product data, including prices, descriptions, user reviews, and ratings.

[1150] The server then aggregates the product data obtained from each e-commerce platform. This aggregated product data is first filtered and narrowed down based on user-specified criteria (e.g., price range, rating, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review rating.

[1151] The ranked product data is sent to the device, which then presents the received product data to the user in a display format such as list view or grid view, allowing users to easily find the best product from a large number of options.

[1152] Finally, the server analyzes the user's past purchasing behavior and search history to provide personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to refine the next search criteria. This allows the user to receive more personalized search results and product suggestions.

[1153] Specific operation example

[1154] For example, if a user wants to find autographed merchandise of a famous athlete and enters the search criteria "I want autographed merchandise of a famous athlete," the system will perform the following actions:

[1155] 1. Getting User Input

[1156] The user types "I want autographed merchandise of a famous athlete" into their device and clicks the "Search" button.

[1157] 2. Analysis of Conditions

[1158] The server passes the received input to a generative AI model, which generates a specific search query: "famous athlete signed merchandise."

[1159] 3. Crawling e-commerce sites

[1160] The server uses the generated search query to crawl multiple e-commerce platforms, such as Amazon, eBay, and Rakuten, to collect relevant product data.

[1161] 4. Filtering and Ranking Results

[1162] The server filters the collected product data based on price, user reviews, etc., and ranks them.

[1163] 5. Providing optimal search results

[1164] The server sends the filtered and ranked product data to the device, which then displays it to the user in a list view, showing each product's image, price, and rating.

[1165] 6. Personalized recommendations and feedback loops

[1166] The server analyzes the products the user clicks on and their purchase history, allowing it to suggest more suitable products for the next search.

[1167] Prompt Sentence Examples

[1168] As an example of a prompt statement, enter the following:

[1169] If a user types in "I want autographed merchandise of a famous athlete," analyze what kind of search query this translates into.

[1170] As described above, the present invention realizes a system that enables a user to efficiently find a desired product even if the user has vague conditions.

[1171] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1172] Step 1:

[1173] A user uses their device to enter vague search criteria into the search bar, such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[1174] Specifically, the user types text into an input field and performs a submit operation.

[1175] Input: Vague conditions entered by the user (e.g., "I want autographed merchandise by a famous athlete")

[1176] Output: Data when the send button is pressed and the search conditions are confirmed

[1177] Step 2:

[1178] After receiving the information entered by the user, the device converts the data into an appropriate data format, such as JSON, and sends it to the server. For example, it converts the data into JSON data such as "{'query': 'I want autographed merchandise of a famous athlete'}".

[1179] Specifically, the device converts the input text into a JSON string and sends it to the server as an HTTP request.

[1180] Input: User input data (e.g., "I want autographed merchandise from a famous athlete")

[1181] Output: Search criteria data in JSON format (e.g., "{'query': 'I want autographed merchandise by famous athletes'}")

[1182] Step 3:

[1183] The server receives the vague search terms sent from the device and passes this data to the generative AI model, which then analyzes the user's vague search terms and converts them into a specific search query (e.g., "famous athlete signed merchandise").

[1184] Specifically, the server calls an API endpoint and passes the data to the generative AI model for analysis.

[1185] Input: Search criteria data in JSON format (e.g., "{'query': 'I want autographed merchandise by famous athletes'}")

[1186] Output: A specific search query (e.g., "famous athlete signed merchandise")

[1187] Step 4:

[1188] The server crawls multiple e-commerce platforms using the search query generated by the generative AI model, sending the search query to each platform and retrieving relevant product data (e.g., price, description, user reviews, ratings, etc.).

[1189] Specifically, the server runs a crawling script and sends API requests to each platform to collect product data.

[1190] Input: A specific search query (e.g., "famous athlete signed merchandise")

[1191] Output: Product data obtained from each platform (e.g. price, description, user reviews, ratings)

[1192] Step 5:

[1193] The server aggregates product data obtained from each e-commerce platform. The aggregated product data is then filtered and narrowed down based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[1194] Specifically, the server executes code to sort and filter the product data.

[1195] Input: Raw data from each platform (e.g. price, description, user reviews, ratings)

[1196] Output: Filtered and ranked product data

[1197] Step 6:

[1198] The server transmits the filtered and ranked product data to the terminal, which presents the received product data to the user in a display format such as a list view or a grid view.

[1199] Specifically, the product list configured on the server side is sent to the terminal as an HTTP response, and the data is displayed as a UI on the terminal side.

[1200] Input: Filtered and ranked product data

[1201] Output: A list of products displayed in list view or grid view

[1202] Step 7:

[1203] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to improve the accuracy of the next search criteria. This provides the user with more personalized search results and product suggestions.

[1204] Specifically, the server accesses the user database and runs an algorithm that analyzes past data.

[1205] Input: User's past purchasing behavior and search history

[1206] Output: personalized product recommendations

[1207] (Application example 1)

[1208] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1209] Conventional e-commerce systems have the problem that it is difficult for users to find the desired product unless they enter precise keywords. Furthermore, searches using vague conditions often result in the display of many irrelevant products, making them inefficient for users. Meanwhile, individual suggestions based on purchase and search history are not fully utilized, making it difficult to suggest optimal products tailored to the user's needs. Furthermore, the limited options for displayed products and the lack of feedback reflected in the next search also detract from the user experience.

[1210] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1211] In this invention, the server includes a means for receiving vague search terms entered by a user, analyzing them based on a generative AI model, and converting them into an appropriate search query; a means for crawling multiple e-commerce platforms and aggregating product data based on the converted search query; and a means for filtering the aggregated product data and ranking them based on the user's specified search terms. This allows users to search for optimal products even with vague search terms. The server also uses the generative AI model to analyze past purchasing behavior and make personalized product suggestions, thereby providing more appropriate products to users. Additionally, the system includes a means for the terminal to allow users to select additional products based on the displayed search results, and a means for sending user feedback to the server to create a feedback loop to improve the accuracy of the next search, thereby improving the user experience.

[1212] "User input" refers to the action of a user providing their desired conditions or requests through an electronic device.

[1213] A "generative AI model" is an algorithm or system that uses artificial intelligence technology to convert vague conditions into specific search queries.

[1214] A "search query" is a keyword or phrase that a user enters into a search engine to retrieve specific information.

[1215] "E-commerce platform" refers to a website or application for buying and selling products and services over the Internet.

[1216] "Crawling" is the process of automatically visiting web pages on the Internet and collecting data.

[1217] "Product data" is a data set that contains detailed product information, such as price, description, and reviews.

[1218] "Filtering" is a process for narrowing down data based on specific conditions.

[1219] "Ranking" is the process of ranking data based on specific criteria.

[1220] "Personalized suggestions" analyze a user's past behavior and preferences to suggest products and services that are best suited to each individual user.

[1221] A "feedback loop" is a system that collects user reactions and behavior and reflects them in the next service provision.

[1222] The system for realizing this invention mainly includes a terminal that processes user input, a server that analyzes ambiguous conditions and generates optimal search queries, and a crawling function that aggregates product data from e-commerce platforms based on the generated queries.

[1223] User input via terminal

[1224] The user inputs their desired conditions and requests into the terminal. For example, the user might input "I want autographed merchandise of a famous athlete." Once the input is confirmed, the terminal sends this information to the server. The terminal can be a smartphone, tablet, PC, or other device.

[1225] Server-based analysis of ambiguous conditions

[1226] The server passes the vague conditions received from the device to a generative AI model. For example, the BART transformer model is used as the generative AI model. The model converts the vague conditions into a specific search query. For example, the condition "I want signed merchandise by a famous athlete" is converted to "famous athlete signed merchandise."

[1227] Crawling product data from e-commerce sites

[1228] The server uses the generated search query to crawl multiple e-commerce platforms (e.g., Amazon and Rakuten) to retrieve relevant product data, including information such as price, description, reviews, and ratings.

[1229] Filtering and Ranking Data

[1230] The acquired product data is filtered on the server and ranked based on user-specified criteria (price range, ratings, etc.) The filtering and ranking algorithms are implemented in Python, for example.

[1231] Providing optimal search results

[1232] The server sends the filtered and ranked product data to the device, which displays it to the user, who can then browse the best products in a list or grid view and select specific products.

[1233] Personalized product recommendations

[1234] The server analyzes the user's past purchasing behavior and search history and uses a generative AI model to make personalized suggestions, thereby providing users with more relevant and personalized product suggestions.

[1235] Formation of a feedback loop

[1236] The device sends user operation information and feedback to the server, which then analyzes the feedback and creates a feedback loop to improve the accuracy of the next search.

[1237] Specific examples

[1238] For example, if a user types in "latest smartphone case recommendations," the server converts this to "latest smartphone cases recommendations," retrieves smartphone case information from Amazon and Rakuten, and displays a list.

[1239] Prompt Sentence Examples

[1240] "Generate what query would be converted when a user types 'I want autographed merchandise of a famous athlete' into a product search app."

[1241] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1242] Step 1:

[1243] Get the conditions entered by the user

[1244] The device accepts input from the user. Specifically, the user enters vague keywords such as "I want autographed merchandise of a famous athlete" into the product search bar. Once this input is confirmed, the device converts the information into a data format such as JSON and sends the data to the server.

[1245] Input: Vague keywords entered by the user

[1246] Output: Keyword information converted into data format such as JSON

[1247] Step 2:

[1248] Convert vague conditions into specific search queries

[1249] The server passes the vague conditions received from the device to a generative AI model. The generative AI model (e.g., the BART transformer model) analyzes the conditions and converts them into specific search queries. For example, "I want autographed merchandise by a famous athlete" is converted to "famous athlete signed merchandise."

[1250] Input: Vague keywords in JSON format sent from the terminal

[1251] Output: A specific search query transformed by the generative AI model

[1252] Step 3:

[1253] Crawl multiple e-commerce platforms based on a search query

[1254] The server uses the generated search query to access the APIs of multiple e-commerce platforms (e.g., Amazon and Rakuten) and crawl product data, including product names, prices, descriptions, reviews, ratings, etc.

[1255] Input: A specific search query transformed by a generative AI model

[1256] Output: Product data obtained from multiple e-commerce platforms

[1257] Step 4:

[1258] Filter and rank your product data

[1259] The server aggregates the acquired product data and filters it based on the user's specified conditions (price range, ratings, etc.), and then ranks the filtered product data based on criteria such as ratings and price.

[1260] Input: Product data obtained from multiple e-commerce platforms

[1261] Output: Filtered and ranked product data

[1262] Step 5:

[1263] Displaying the most relevant product information to users

[1264] The server transmits the filtered and ranked product data to the terminal, which displays the data to the user in a format such as a list view or a grid view, allowing the user to review and select products.

[1265] Input: Filtered and ranked product data

[1266] Output: Product information displayed in list or grid view format

[1267] Step 6:

[1268] Make personalized product recommendations based on past purchasing behavior and search history

[1269] The server analyzes the user's past purchasing behavior and search history, and uses a generative AI model based on that information to make personalized product suggestions, allowing users to be recommended products that match their preferences.

[1270] Input: User's past purchasing behavior and search history

[1271] Output: personalized product recommendations

[1272] Step 7:

[1273] Forming a feedback loop

[1274] The device sends user operation information and feedback to the server, which analyzes this feedback and reflects it in order to improve the accuracy of the next search, allowing the user to obtain more accurate search results.

[1275] Input: User operation information and feedback

[1276] Output: Data reflected to improve search accuracy next time

[1277] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1278] In an embodiment of the present invention, a system is provided that allows users to easily search for desired products using vague search criteria and obtain optimal results. Furthermore, the system recognizes the user's emotions and improves the accuracy of search queries based on those emotions, thereby providing more personalized product offers. A specific embodiment of the system and its operation are described below.

[1279] Getting User Input

[1280] User

[1281] Users enter vague criteria into the search bar on their device (smartphone, tablet, PC, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." Once they've finished entering the information, they click the "Search" button.

[1282] Terminal

[1283] The terminal receives the information entered by the user, converts the data into an appropriate data format such as JSON, and sends it to the server.

[1284] Condition analysis and emotion recognition

[1285] server

[1286] The server receives the ambiguous condition data sent from the device, passes this data to the generative AI model, and begins the analysis process.

[1287] Generative AI Models

[1288] The generative AI model analyzes the user's vague search criteria and converts them into a specific search query, which is then used to provide optimal search results.

[1289] Emotion Engine

[1290] The emotion engine analyzes the user's emotions at the time of input from voice, text input, facial expressions, etc. Based on the results of this analysis, the generative AI model optimizes its condition analysis. For example, if the user is excited, it will make more proactive product suggestions, while if the user is calm, it will emphasize more detailed information.

[1291] Crawling e-commerce sites

[1292] server

[1293] The server uses the generated search query and the sentiment data obtained from the sentiment engine to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.) and sends a search query to each platform to retrieve related product data.

[1294] Filtering and ranking results

[1295] server

[1296] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. This dataset is then filtered to narrow down the list based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[1297] Providing optimal search results

[1298] server

[1299] The server transmits the filtered and ranked product data to the terminal.

[1300] Terminal

[1301] The device displays the received product data to the user in a list view, grid view, or other display format, allowing the user to easily find the best product from a wide range of options.

[1302] Personalized recommendations and feedback loops

[1303] server

[1304] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is stored and used to improve the accuracy of the next search criteria. User emotional data is also stored and reflected in future searches and suggestions.

[1305] user

[1306] Users can select a product from the product list displayed on their device and proceed to the details page or purchase page. They can also enter new search criteria and search again.

[1307] Specific examples

[1308] For example, suppose a user who wants to find autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete" and seems to be having fun. In this case, the system will perform the following actions.

[1309] 1. User Input

[1310] The user types "I want autographed merchandise from a famous athlete" into the device.

[1311] 2. Condition Analysis and Emotion Recognition

[1312] The server passes the input to a generative AI model for analysis, resulting in the search query "famous athlete signed merchandise."

[1313] The emotion engine analyzes the user's positive emotions and, based on the generated query, prioritizes products with more positive reviews.

[1314] 3. E-commerce site crawling

[1315] The server crawls multiple e-commerce platforms and collects product data.

[1316] 4. Filtering and Ranking Results

[1317] The server filters the collected product data and prioritizes and ranks products that have high ratings and many positive reviews that reflect the user's enjoyment emotions.

[1318] 5. Search result presentation

[1319] The server sends the filtered and ranked results to the device.

[1320] The device displays the appropriate product list to the user.

[1321] 6. Personalized Offers

[1322] The server analyzes the user's behavioral and emotional data to make more relevant personalized suggestions in the future.

[1323] In this way, a system is provided that allows users to easily find the products they are looking for without knowing the specific product names or detailed conditions, and furthermore, allows them to receive appropriate product suggestions based on their emotions.

[1324] The processing flow will be explained below.

[1325] Step 1:

[1326] User: Enters vague criteria such as "I want autographed merchandise of a famous athlete" into the device's search bar. Once entered, presses the "Search" button.

[1327] Step 2:

[1328] Terminal: Receives the search criteria entered by the user, converts the data into an appropriate format (e.g., JSON), and sends it to the server. An example of sending data is {"query":"I want autographed merchandise by a famous athlete"}.

[1329] Step 3:

[1330] Server: Receives search criteria data sent from the device and temporarily stores this data in memory.

[1331] Step 4:

[1332] Server: Passes the received data to the generative AI model and begins the analysis process. The generative AI model analyzes the user's vague search criteria and converts them into an appropriate search query (e.g., "famous athlete signed merchandise").

[1333] Step 5:

[1334] Emotion engine: Analyzes the emotions at the time of input from the user's voice, text input, facial expressions, etc. The emotion engine outputs the analysis results and optimizes the condition analysis of the generative AI model.

[1335] Step 6:

[1336] Server: Stores the search queries generated by the generative AI model and the sentiment analysis data from the sentiment engine, and prepares them to be sent to multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.).

[1337] Step 7:

[1338] Server: Uses generated search queries to access APIs and crawl the web for each e-commerce platform, and collects product data (price, description, ratings, etc.) from each platform.

[1339] Step 8:

[1340] Each e-commerce site: Returns a list of products based on the search query to the server through the provided API or web interface. For example, the results are returned in the format {"site": "Amazon", "results": [...]}.

[1341] Step 9:

[1342] Server: Aggregates product data collected from each e-commerce platform to form a large product dataset, which is temporarily stored in memory.

[1343] Step 10:

[1344] Server: Filters the aggregated product data. Based on the user's criteria (price range, ratings, etc.), it runs a process to narrow down the list to only products that match the criteria. It also prioritizes products based on sentiment data (e.g., if the user seems happy, prioritize products with many positive reviews).

[1345] Step 11:

[1346] Server: Ranks the filtered products. Ranking criteria include price, popularity, review rating, etc. Generates a ranked product list.

[1347] Step 12:

[1348] Server: Sends the filtered and ranked product list to the device. Example data sent includes {"ranked_products": [...]}.

[1349] Step 13:

[1350] Device: The received product list is displayed to the user in a format appropriate for the user interface, such as list view or grid view.

[1351] Step 14:

[1352] Users can select the product they are interested in from the list of products presented and proceed to the detail page or purchase page. If they do not like the product, they can enter new search criteria and search again.

[1353] Step 15:

[1354] Server: Records user behavior (selection, confirmation of details, purchase, etc.) and emotional data, and stores it as analytical data to improve search accuracy next time.

[1355] Step 16:

[1356] Server: Prepares to make future personalized product recommendations based on the analyzed user and sentiment data.

[1357] This series of steps allows users to easily find the products they are looking for and allows the system to continuously improve its accuracy.

[1358] Example 2

[1359] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1360] In conventional e-commerce platforms, when users search for products using vague criteria, they often encounter problems such as insufficient search results or being unable to find the desired product. Furthermore, because the platform does not take into account the user's purchasing motivation or emotions, personalized product suggestions are not possible, often resulting in low satisfaction. The present invention aims to solve these problems and provide users with optimal product search results and personalized suggestions.

[1361] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1362] In this invention, the server includes a means for analyzing vague search terms entered by a user based on a generative AI model and converting them into an appropriate search query, an emotion recognition means for analyzing the user's emotional state, and a means for crawling multiple e-commerce platforms and aggregating product data based on the converted search query. This allows a specific search query appropriate to the user's intent to be generated, even if the user enters vague search terms, and the aggregated product data can be ranked based on the user's specified search terms and emotional state. Furthermore, the system analyzes the user's past behavioral data and emotional data to provide personalized product suggestions, thereby increasing user satisfaction.

[1363] "Ambiguous search conditions entered by the user" refers to non-specific search conditions entered by the user through the terminal.

[1364] A "generative AI model" refers to an artificial intelligence model that analyzes vague conditions entered by users and converts them into appropriate search queries.

[1365] A "search query" refers to a specific search term or phrase that a generative AI model generates by analyzing ambiguous conditions.

[1366] "Emotion recognition means" refers to technology that analyzes the emotional state of a user from their input or interaction, including, for example, voice, text input, and facial expression recognition.

[1367] An "e-commerce platform" is a website or application that enables the buying and selling of goods and services online, such as Amazon or eBay.

[1368] "Crawling" refers to the act of a server automatically searching for information on a particular website and collecting data.

[1369] "Product Data" refers to all product-related information obtained from the e-commerce platform, such as product name, price, review rating, etc.

[1370] "Filtering" refers to the process of narrowing down collected product data based on user-specified conditions.

[1371] "Ranking" refers to the process of ranking filtered product data based on specific criteria, such as price, popularity, or review rating.

[1372] "Specified conditions" refer to conditions that a user sets to narrow down search results, such as a price range or rating range.

[1373] "Individualized product proposals" refer to product proposals that are optimized for each individual user based on the user's past behavioral and emotional data.

[1374] A "feedback loop" refers to a system that receives feedback from users and uses that information to improve the accuracy of subsequent searches and suggestions.

[1375] The present invention is a system for providing optimal search results when a user searches for products under vague conditions. Furthermore, the system recognizes the user's emotions and improves the accuracy of the search query based on the emotions, thereby providing more personalized product suggestions. A specific embodiment of the system and its operation are described below.

[1376] 1. Getting User Input

[1377] User

[1378] Users enter vague criteria into the search bar on their device (smartphone, tablet, PC, etc.), for example, an abstract request such as "I want autographed merchandise of a famous athlete." Once they're done, they click the "Search" button.

[1379] Terminal

[1380] The device acquires the conditions entered by the user, converts this input data into an appropriate data format such as JSON, and sends it to the server. At this time, when the user has completed entering the information on the device, the device sends an HTTP request to the server.

[1381] 2. Condition analysis and emotion recognition

[1382] server

[1383] The server receives vague condition data sent from the device. This data is passed to the generative AI model as a prompt. For example, the data {"query": "I want autographed merchandise of a famous athlete"} is sent.

[1384] Generative AI Models

[1385] The generative AI model analyzes the received prompt and converts the user's vague search criteria into a specific search query, such as "famous athlete signed merchandise." This converted query is cached on the server and can be reused.

[1386] Emotion Engine

[1387] The emotion engine analyzes user input (voice, text, facial expressions, etc.) to recognize the user's emotional state. For example, facial recognition technology can be used to determine whether a user is happy when typing. This emotional data influences the analysis of the generative AI model.

[1388] 3. Crawling e-commerce sites

[1389] server

[1390] The server uses the converted search query and sentiment data to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.). For each platform, it sends a new search query and collects related product data. For example, it uses the Amazon API to send the query "famous athlete signed merchandise" and retrieves the corresponding product data.

[1391] 4. Filtering and Ranking Results

[1392] server

[1393] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. This dataset is then filtered to narrow down the results based on user-specified criteria (price range, ratings, etc.). Specifically, it selects highly rated products within a certain price range. It also prioritizes products with proactive product suggestions and detailed information based on emotional data. For example, for users with a positive emotional state, it prioritizes products with many positive reviews.

[1394] 5. Providing optimal search results

[1395] server

[1396] The server sends the filtered and ranked product data to the terminal, for example, returning a product list in JSON format.

[1397] Terminal

[1398] The device displays the received product data to the user. The display format can be selected from list view, grid view, etc. For example, in list view, the product name, image, price, and review rating are displayed.

[1399] 6. Personalized recommendations and feedback loops

[1400] server

[1401] The server analyzes users' past purchasing behavior and search history to make personalized product suggestions. It also accumulates behavioral data, such as when users click on a specific product to view its details or make a purchase. This data is analyzed and saved along with emotional data to help improve the accuracy of the next search.

[1402] User

[1403] Users can select products they are interested in from the products displayed on their devices, proceed to the details and purchase process, or enter new search criteria and search again.

[1404] Specific examples

[1405] For example, suppose a user searching for autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete" and seems to be having fun. In this case, the system performs the following specific actions:

[1406] 1. User Input

[1407] The user types "I want autographed merchandise from a famous athlete" into the device.

[1408] The user clicks the Search button.

[1409] 2. Condition Analysis and Emotion Recognition

[1410] The server passes the input data to the generative AI model.

[1411] A generative AI model generates the search query "famous athlete signed merchandise."

[1412] The emotion engine analyzes the user's happy facial expression.

[1413] 3. E-commerce site crawling

[1414] A server sends search queries to multiple e-commerce platforms to collect product data.

[1415] 4. Filtering and Ranking Results

[1416] The server filters the collected product data and prioritizes products with positive reviews and high ratings.

[1417] 5. Search result presentation

[1418] The server transmits the ranked product data to the terminal.

[1419] The device displays the product list to the user in a list view format.

[1420] 6. Personalized Offers

[1421] The server analyzes the user's behavioral and emotional data to make more appropriate personalized suggestions in the future.

[1422] Example prompt sentence:

[1423] "I want autographed merchandise from famous athletes."

[1424] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1425] Step 1:

[1426] Getting User Input

[1427] The user enters vague search criteria into the search bar of their device (smartphone, tablet, PC, etc.). For example, they might type "I want autographed merchandise of a famous athlete." Then, they click the "Search" button. The input data is in text format.

[1428] Step 2:

[1429] Sending data from the device to the server

[1430] The device converts the user-entered criteria into an appropriate data format, such as JSON. The device then sends this data to the server using an HTTP request, such as {"query": "I want autographed merchandise by a famous athlete"}. This data is provided by the user as input.

[1431] Step 3:

[1432] Condition Analysis

[1433] The server passes the vague condition data received from the device to the generative AI model. The generative AI model analyzes this data and converts it into a specific search query. For example, the vague condition "I want signed merchandise by a famous athlete" is converted into the specific search query "famous athlete signed merchandise." This converted query is output as search data.

[1434] Step 4:

[1435] Emotion Analysis

[1436] The server uses an emotion engine to analyze the user's input format (voice, text, facial expression, etc.). The emotion engine analyzes emotions from the user's facial expression and voice data at the time of input and returns the results to the server. For example, facial expression analysis can yield an emotion result of "looking happy." This emotion data becomes input for use in optimizing search queries.

[1437] Step 5:

[1438] Crawling e-commerce sites

[1439] The server uses the generated search query and the analyzed sentiment data to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.). During the crawling, it uses the platform's API or scraping technology to send the search query and retrieve the corresponding product data. For example, it sends the search query "famous athlete signed merchandise" to Amazon to retrieve a list of related products. This product data is output as a result from the e-commerce site.

[1440] Step 6:

[1441] Filtering and ranking results

[1442] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. Products are filtered from this dataset based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on emotion data. For example, products with many positive reviews are prioritized for users with positive emotions. The filtered and ranked product data is output as an optimal product list.

[1443] Step 7:

[1444] Providing optimal search results

[1445] The server sends the filtered and ranked product data to the device, for example in the format {"products": [listed product data]}. This data is returned to the user as the best search results.

[1446] Terminal

[1447] The device displays the received product data to the user. The display format can be selected from list view, grid view, etc. For example, list view displays the product name, image, price, and review rating. This displayed product information is provided in a format that is easy for the user to view.

[1448] Step 8:

[1449] Personalized recommendations and feedback loops

[1450] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. It also accumulates behavioral data, such as when a user clicks on a specific product to view its details page or make a purchase. This data is analyzed to help improve the accuracy of the next search criteria. For example, if a user repeatedly takes positive action regarding "signature goods," similar products will be included in the next suggestion. This feedback data is output as the user's behavioral pattern.

[1451] User

[1452] Users can select products they are interested in from the displayed products and proceed to the purchase process. They can also search again by entering new search criteria, improving the user experience. These new search criteria are then processed as data to be input again into the system.

[1453] As described above, the system of the present invention can efficiently carry out a series of processes from vague conditions to specific product proposals.

[1454] (Application example 2)

[1455] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1456] Existing search systems have difficulty finding appropriate search results for vague user inputs, making it difficult for users to smoothly find the products they are looking for. Furthermore, the lack of personalized suggestions based on user emotions and past behavior makes it difficult to provide a more personalized shopping experience. To solve these problems, there is a need for a search system that takes user emotions into account.

[1457] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving ambiguous conditions entered by a user and transmitting them to the server, means for the server to analyze the ambiguous conditions received based on a generative AI model and convert them into an appropriate search query, means for the server to crawl multiple e-commerce platforms and aggregate product data based on the converted search query, means for the server to filter the aggregated product data using an emotion recognition engine taking into account the user's emotional data and rank it based on the user's specified conditions, and means for the server to transmit the filtered and ranked product data to a terminal, which then displays it to the user. This enables users to accurately search for products even with ambiguous conditions and receive more personalized product suggestions based on their emotions.

[1458] "User" means an individual or organization that uses the System to search for and purchase products.

[1459] "Vague terms" are abstract search queries or requests entered by users without specifying specific products or specifications.

[1460] "Server" refers to a computer system that receives input data from users and performs processes such as analysis using generative AI models, emotion recognition engines, crawling e-commerce platforms, filtering and ranking data, etc.

[1461] A "generative AI model" is a type of artificial intelligence that analyzes ambiguous conditions and converts them into appropriate search queries, and is based on natural language processing.

[1462] An "emotion recognition engine" is a technology for analyzing emotions from non-verbal information such as a user's facial expressions and voice.

[1463] An "e-commerce platform" is an online marketplace for buying and selling goods and services over the Internet.

[1464] "Crawling" is the process of automatically gathering information on a website.

[1465] "Filtering" is the process of removing unnecessary data from collected product data based on specific criteria and extracting only the necessary data.

[1466] "Ranking" is the process of ranking the filtered product data based on specific evaluation criteria.

[1467] A "terminal" is a device that a user uses to access the system, including smartphones and computers.

[1468] A "feedback loop" is a cyclical process that analyzes information collected from users (e.g., search history, sentiment data) to improve the accuracy of future search criteria and suggestions.

[1469] The system for implementing this invention allows users to search for products under vague conditions and optimizes the results based on emotions. The specific configuration and operation of this system will be described below.

[1470] System Configuration

[1471] Hardware

[1472] Smartphone: A device with a built-in camera and microphone that allows users to access the system.

[1473] Server: A computer system equipped with a high-speed processor and capable of processing large amounts of data. Cloud-based operation is recommended.

[1474] software

[1475] Generative AI model: A natural language processing model to convert users' vague search terms into specific search queries. In this case, we use OpenAI's GPT-4.

[1476] Emotion recognition engine: An engine that analyzes the user's facial expressions and voice to recognize emotions. It uses Google Cloud Vision API and Microsoft Azure Face API.

[1477] E-commerce site crawling library: A library for collecting product data from e-commerce platforms. It uses BeautifulSoup (Python) and other tools.

[1478] Cloud platform: Infrastructure for operations and data processing. We recommend using Amazon Web Services (AWS).

[1479] System Operation

[1480] 1. Getting User Input

[1481] Users enter vague terms into the search bar on their smartphone and press the search button.

[1482] Example: "I want new shoes."

[1483] 2. Emotion recognition

[1484] The device's camera and microphone are used to collect the user's facial expressions and voice.

[1485] The emotion recognition engine analyzes this data and recognizes the user's emotions.

[1486] Example: A user looks happy.

[1487] 3. Condition analysis

[1488] Vague search queries are fed into a generative AI model, which converts them into specific search queries.

[1489] Example: "new shoes"

[1490] 4. Crawling e-commerce sites

[1491] The server uses the generated search query to crawl multiple e-commerce platforms to collect product data.

[1492] The collected data is converted into a unified format and aggregated on the server.

[1493] 5. Filtering and Ranking Results

[1494] The aggregated product data is filtered taking into account user sentiment data.

[1495] The filtered data is then ranked based on criteria such as price, popularity, and review ratings.

[1496] 6. Presentation of search results

[1497] The server transmits the filtered and ranked data to the terminal.

[1498] The terminal displays a list of products that are most suitable for the user.

[1499] 7. Creating feedback loops

[1500] User selections, search history, emotional data, etc. are stored on the server to improve search accuracy from the next time onwards.

[1501] Prompt Sentence Examples

[1502] Here is an example of a prompt for the generative AI model:

[1503] A user searched for: "I want new shoes." The user's sentiment seems to be happy. Convert this into a specific search query.

[1504] In this way, users can easily find the products they are looking for without having to enter specific product names or detailed search criteria. Furthermore, the system's emotion-based product suggestions enable a more personalized shopping experience.

[1505] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1506] Step 1: Getting User Input

[1507] The user enters vague search criteria into the device, for example, "I want new shoes." Once the search is complete, the user presses the search button. The device converts the input data into JSON format and sends it to the server.

[1508] Input: User's vague search criteria

[1509] Output: Input data in JSON format

[1510] Specific actions: Enter search criteria in the search bar and press the button.

[1511] Step 2: Emotion Recognition

[1512] The device uses a built-in camera and microphone to collect the user's facial expressions and voice. This data is sent to an emotion recognition engine to analyze the user's emotions. The analysis results are sent to the server in JSON format.

[1513] Input: User's facial expression and voice data

[1514] Output: Emotion data in JSON format

[1515] Specific actions: Capture facial expressions and voice using a camera and microphone.

[1516] Step 3: Condition analysis

[1517] The server receives the search criteria and emotion data sent from the device, passes this data to the generative AI model, and converts it into a specific search query. The generative AI model analyzes the prompt and generates an appropriate search query.

[1518] Input: Fuzzy search terms and sentiment data in JSON format

[1519] Output: Specific search query

[1520] Specific operation: The generative AI model analyzes the prompt sentence and generates a search query.

[1521] Step 4: Crawl your e-commerce site

[1522] The server crawls multiple e-commerce platforms using the generated specific search query, collects product data using libraries such as BeautifulSoup, converts the collected data into a unified format, and aggregates it on the server.

[1523] Input: A specific search query

[1524] Output: Aggregated product data in a unified format

[1525] Specific operation: Collects product data using e-commerce site crawl libraries.

[1526] Step 5: Filter and rank the results

[1527] The server receives the aggregated product data and filters it using an emotion recognition engine, taking into account the user's emotional data. The filtered data is then ranked based on criteria such as price, ratings, and reviews.

[1528] Input: Aggregated product data and sentiment data in a unified format

[1529] Output: Filtered and ranked product data

[1530] Specific behavior: Analyze product data and optimize based on user sentiment.

[1531] Step 6: Presenting search results

[1532] The server sends the filtered and ranked product data to the device, which then visually displays this data to the user in a format such as a list view or grid view.

[1533] Input: Filtered and ranked product data

[1534] Output: The optimal product list shown to the user

[1535] Specific operation: Display the product list on the terminal screen.

[1536] Step 7: Creating a feedback loop

[1537] The server stores user behavioral history and emotional data to improve the accuracy of search criteria and suggestions from the next time onward, allowing users to receive more personalized search results.

[1538] Input: User behavior history, emotional data

[1539] Output: Feedback to improve accuracy using accumulated data

[1540] Specific actions: Accumulate behavioral history and emotional data and reflect them in the next proposal.

[1541] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1542] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1543] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1544] [Fourth embodiment]

[1545] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1546] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1547] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1548] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1549] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1550] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1551] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1552] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1553] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1554] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1555] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1556] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1557] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1558] In the embodiment of the present invention, a system is provided that allows users to easily search for products they want and obtain optimal results. A specific embodiment of this system and its operation will be described below.

[1559] Getting User Input

[1560] User

[1561] Users enter vague criteria into the search bar on their devices (smartphones, tablets, PCs, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[1562] Terminal

[1563] The terminal accepts information entered by the user, converts the data into an appropriate data format such as JSON, and sends it to the server.

[1564] Condition Analysis

[1565] server

[1566] The server receives vague conditions sent from the device. This data is passed to the generative AI model. The generative AI model analyzes the user's vague conditions and converts them into a specific search query. This converted search query is optimized to search for the product the user is looking for.

[1567] Crawling e-commerce sites

[1568] server

[1569] The server crawls multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.) using the search query generated by the generative AI model. It sends the search query to each platform and retrieves relevant product data, including prices, descriptions, user reviews, and ratings.

[1570] Filtering and ranking results

[1571] server

[1572] The server aggregates product data obtained from each e-commerce platform. The aggregated product data is filtered and narrowed down based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[1573] Providing optimal search results

[1574] server

[1575] The server transmits the filtered and ranked product data to the terminal.

[1576] Terminal

[1577] The device displays the received product data to the user in a list view, grid view, or other display format, allowing the user to easily find the best product from a wide range of options.

[1578] Personalized recommendations and feedback loops

[1579] server

[1580] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to improve the accuracy of the next search criteria. This provides the user with more personalized search results and product suggestions.

[1581] user

[1582] Users can select a product from the product list displayed on their device and proceed to the details page or purchase page. They can also enter new search criteria and search again.

[1583] Specific examples

[1584] For example, if a user who wants to find autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete," the system will perform the following actions.

[1585] 1. User Input

[1586] The user types "I want autographed merchandise from a famous athlete" into the device.

[1587] 2. Condition analysis

[1588] The server passes the input to a generative AI model for analysis, resulting in the search query "famous athlete signed merchandise."

[1589] 3. E-commerce site crawling

[1590] The server crawls multiple e-commerce platforms and collects product data.

[1591] 4. Filtering and Ranking Results

[1592] The server filters the collected product data and ranks it based on criteria such as price and ratings.

[1593] 5. Search result presentation

[1594] The server sends the filtered and ranked results to the device.

[1595] The device displays the appropriate product list to the user.

[1596] 6. Personalized Offers

[1597] The server analyzes user behavior and makes more relevant personalized offers in the future.

[1598] In this way, a system is provided that allows users to easily find the product they are looking for, even if they do not know the specific product name or detailed conditions.

[1599] The processing flow will be explained below.

[1600] Step 1:

[1601] User: Enters vague criteria such as "I want autographed merchandise of a famous athlete" into the device's search bar. Once entered, presses the "Search" button.

[1602] Step 2:

[1603] Terminal: Receives the search criteria entered by the user, converts the data into an appropriate format (e.g., JSON), and sends it to the server. An example of sending data is {"query":"I want autographed merchandise by a famous athlete"}.

[1604] Step 3:

[1605] Server: Receives search criteria data sent from the device and temporarily stores this data in memory.

[1606] Step 4:

[1607] Server: Passes the received data to the generative AI model and begins the analysis process. The generative AI model analyzes the user's vague search criteria and converts them into an appropriate search query (e.g., "famous athlete signed merchandise").

[1608] Step 5:

[1609] Server: Stores the search queries generated by the generative AI model and prepares them to be sent to multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten).

[1610] Step 6:

[1611] Server: Uses generated search queries to access APIs and crawl the web for each e-commerce platform, and collects product data (price, description, ratings, etc.) from each platform.

[1612] Step 7:

[1613] Each e-commerce site: Returns a list of products based on the search query to the server through the provided API or web interface. For example, the results are returned in the format {"site": "Amazon", "results": [...]}.

[1614] Step 8:

[1615] Server: Aggregates product data collected from each e-commerce platform to form a large product dataset, which is temporarily stored in memory.

[1616] Step 9:

[1617] Server: Filters the aggregated product data. Based on the user's specified criteria (price range, ratings, etc.), the process narrows down the results to only those products that match the criteria.

[1618] Step 10:

[1619] Server: Ranks the filtered products. Ranking criteria include price, popularity, review rating, etc. Generates a ranked product list.

[1620] Step 11:

[1621] Server: Sends the filtered and ranked product list to the device. Example data sent includes {"ranked_products": [...]}.

[1622] Step 12:

[1623] Device: The received product list is displayed to the user in a format appropriate for the user interface, such as list view or grid view.

[1624] Step 13:

[1625] Users can select the product they are interested in from the list of products presented and proceed to the detail page or purchase page. If they do not like the product, they can enter new search criteria and search again.

[1626] Step 14:

[1627] Server: Records user actions (selection, confirmation of details, purchase, etc.) and stores them as analytical data to improve search accuracy next time.

[1628] Step 15:

[1629] Server: Prepares to make personalized product suggestions based on the analyzed user data.

[1630] This series of steps allows users to easily find the products they are looking for and allows the system to continuously improve its accuracy.

[1631] Example 1

[1632] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1633] When searching for a wide variety of products on the Internet, modern consumers often spend a lot of time and effort trying to find the product they are looking for based on vague or abstract criteria. Furthermore, existing search engines and e-commerce platforms are unable to fully utilize users' past search history and purchasing behavior, resulting in a lack of personalized and appropriate product suggestions. This makes it difficult for users to efficiently find the products they need.

[1634] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1635] In this invention, the server includes means for receiving vague conditions entered by a user and transmitting them to the server, means for the server to analyze the received vague conditions based on a generative AI model and convert them into an appropriate search query, means for the server to crawl multiple e-commerce platforms and aggregate product data based on the converted search query, means for the server to filter the aggregated product data and rank it based on the user's specified conditions, means for the server to transmit the filtered and ranked product data to a terminal and for the terminal to display it to the user, and means for the terminal to present the received product data to the user in a display format such as a list view or grid view, thereby enabling users to efficiently find desired products even with vague conditions.

[1636] A "user" is a person who enters vague criteria and receives search results.

[1637] "Device" means a device used by a User to enter search criteria and view product data returned by the Server. Examples include smartphones, tablets, and PCs.

[1638] The "server" refers to a central computer that receives vague conditions sent by users, converts them into search queries using a generative AI model, and retrieves the necessary product data from multiple e-commerce platforms.

[1639] A "generative AI model" refers to an artificial intelligence technique that analyzes vague conditions and converts them into specific search queries.

[1640] A "search query" refers to a search condition that converts a user's vague conditions into a specific and optimal form.

[1641] An "e-commerce platform" refers to an online marketplace that offers goods and services over the web. Examples include Amazon, eBay, and Rakuten.

[1642] "Product Data" refers to information retrieved from e-commerce platforms based on a search query, including prices, descriptions, user reviews, and ratings.

[1643] "Filtering" refers to the process of narrowing down the acquired product data based on the user's specified conditions.

[1644] "Ranking" refers to the process of hierarchically arranging filtered product data based on criteria such as price, popularity, and review rating.

[1645] "List view" refers to a method of displaying product data in a vertical list format.

[1646] "Grid view" refers to a method of displaying product data in multiple rows and columns.

[1647] A "feedback loop" is a system that analyzes users' search results and purchasing behavior history and uses them to improve the accuracy of their next search.

[1648] In an embodiment of the present invention, a system is provided that allows users to easily search for desired products and obtain optimal results. This system is composed of a series of processes: acquiring user input, analyzing conditions, crawling e-commerce sites, filtering and ranking the results, presenting optimal search results, and providing personalized suggestions and a feedback loop.

[1649] Hardware and Software Configuration

[1650] User

[1651] Users enter vague criteria into the search bar on their devices (smartphones, tablets, PCs, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[1652] Terminal

[1653] The device receives the information entered by the user. It then converts the data into an appropriate data format, such as JSON, and sends it to the server. For example, the vague query entered by the user, "I want autographed merchandise by a famous athlete," can be converted to "{"query": "I want autographed merchandise by a famous athlete"}".

[1654] server

[1655] The server receives vague conditions sent from the device. This data is passed to a generative AI model (e.g., GPT-3). The generative AI model analyzes the user's vague conditions and converts them into specific search queries (e.g., "famous athlete signed merchandise"). The converted search queries are optimized for crawling e-commerce sites (e.g., Amazon, eBay, Rakuten).

[1656] The server crawls multiple e-commerce platforms using the search query generated by the generative AI model, sends the search query to each platform, and retrieves relevant product data, including prices, descriptions, user reviews, and ratings.

[1657] The server then aggregates the product data obtained from each e-commerce platform. This aggregated product data is first filtered and narrowed down based on user-specified criteria (e.g., price range, rating, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review rating.

[1658] The ranked product data is sent to the device, which then presents the received product data to the user in a display format such as list view or grid view, allowing users to easily find the best product from a large number of options.

[1659] Finally, the server analyzes the user's past purchasing behavior and search history to provide personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to refine the next search criteria. This allows the user to receive more personalized search results and product suggestions.

[1660] Specific operation example

[1661] For example, if a user wants to find autographed merchandise of a famous athlete and enters the search criteria "I want autographed merchandise of a famous athlete," the system will perform the following actions:

[1662] 1. Getting User Input

[1663] The user types "I want autographed merchandise of a famous athlete" into their device and clicks the "Search" button.

[1664] 2. Analysis of Conditions

[1665] The server passes the received input to a generative AI model, which generates a specific search query: "famous athlete signed merchandise."

[1666] 3. Crawling e-commerce sites

[1667] The server uses the generated search query to crawl multiple e-commerce platforms, such as Amazon, eBay, and Rakuten, to collect relevant product data.

[1668] 4. Filtering and Ranking Results

[1669] The server filters the collected product data based on price, user reviews, etc., and ranks them.

[1670] 5. Providing optimal search results

[1671] The server sends the filtered and ranked product data to the device, which then displays it to the user in a list view, showing each product's image, price, and rating.

[1672] 6. Personalized recommendations and feedback loops

[1673] The server analyzes the products the user clicks on and their purchase history, allowing it to suggest more suitable products for the next search.

[1674] Prompt Sentence Examples

[1675] As an example of a prompt statement, enter the following:

[1676] If a user types in "I want autographed merchandise of a famous athlete," analyze what kind of search query this translates into.

[1677] As described above, the present invention realizes a system that enables a user to efficiently find a desired product even if the user has vague conditions.

[1678] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1679] Step 1:

[1680] A user uses their device to enter vague search criteria into the search bar, such as "I want autographed merchandise of a famous athlete." This input is confirmed by clicking the "Search" button or pressing the Enter key.

[1681] Specifically, the user types text into an input field and performs a submit operation.

[1682] Input: Vague conditions entered by the user (e.g., "I want autographed merchandise by a famous athlete")

[1683] Output: Data when the send button is pressed and the search conditions are confirmed

[1684] Step 2:

[1685] After receiving the information entered by the user, the device converts the data into an appropriate data format, such as JSON, and sends it to the server. For example, it converts the data into JSON data such as "{'query': 'I want autographed merchandise of a famous athlete'}".

[1686] Specifically, the device converts the input text into a JSON string and sends it to the server as an HTTP request.

[1687] Input: User input data (e.g., "I want autographed merchandise from a famous athlete")

[1688] Output: Search criteria data in JSON format (e.g., "{'query': 'I want autographed merchandise by famous athletes'}")

[1689] Step 3:

[1690] The server receives the vague search terms sent from the device and passes this data to the generative AI model, which then analyzes the user's vague search terms and converts them into a specific search query (e.g., "famous athlete signed merchandise").

[1691] Specifically, the server calls an API endpoint and passes the data to the generative AI model for analysis.

[1692] Input: Search criteria data in JSON format (e.g., "{'query': 'I want autographed merchandise by famous athletes'}")

[1693] Output: A specific search query (e.g., "famous athlete signed merchandise")

[1694] Step 4:

[1695] The server crawls multiple e-commerce platforms using the search query generated by the generative AI model, sending the search query to each platform and retrieving relevant product data (e.g., price, description, user reviews, ratings, etc.).

[1696] Specifically, the server runs a crawling script and sends API requests to each platform to collect product data.

[1697] Input: A specific search query (e.g., "famous athlete signed merchandise")

[1698] Output: Product data obtained from each platform (e.g. price, description, user reviews, ratings)

[1699] Step 5:

[1700] The server aggregates product data obtained from each e-commerce platform. The aggregated product data is then filtered and narrowed down based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[1701] Specifically, the server executes code to sort and filter the product data.

[1702] Input: Raw data from each platform (e.g. price, description, user reviews, ratings)

[1703] Output: Filtered and ranked product data

[1704] Step 6:

[1705] The server transmits the filtered and ranked product data to the terminal, which presents the received product data to the user in a display format such as a list view or a grid view.

[1706] Specifically, the product list configured on the server side is sent to the terminal as an HTTP response, and the data is displayed as a UI on the terminal side.

[1707] Input: Filtered and ranked product data

[1708] Output: A list of products displayed in list view or grid view

[1709] Step 7:

[1710] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is accumulated and used to improve the accuracy of the next search criteria. This provides the user with more personalized search results and product suggestions.

[1711] Specifically, the server accesses the user database and runs an algorithm that analyzes past data.

[1712] Input: User's past purchasing behavior and search history

[1713] Output: personalized product recommendations

[1714] (Application example 1)

[1715] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1716] Conventional e-commerce systems have the problem that it is difficult for users to find the desired product unless they enter precise keywords. Furthermore, searches using vague conditions often result in the display of many irrelevant products, making them inefficient for users. Meanwhile, individual suggestions based on purchase and search history are not fully utilized, making it difficult to suggest optimal products tailored to the user's needs. Furthermore, the limited options for displayed products and the lack of feedback reflected in the next search also detract from the user experience.

[1717] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1718] In this invention, the server includes a means for receiving vague search terms entered by a user, analyzing them based on a generative AI model, and converting them into an appropriate search query; a means for crawling multiple e-commerce platforms and aggregating product data based on the converted search query; and a means for filtering the aggregated product data and ranking them based on the user's specified search terms. This allows users to search for optimal products even with vague search terms. The server also uses the generative AI model to analyze past purchasing behavior and make personalized product suggestions, thereby providing more appropriate products to users. Additionally, the system includes a means for the terminal to allow users to select additional products based on the displayed search results, and a means for sending user feedback to the server to create a feedback loop to improve the accuracy of the next search, thereby improving the user experience.

[1719] "User input" refers to the action of a user providing their desired conditions or requests through an electronic device.

[1720] A "generative AI model" is an algorithm or system that uses artificial intelligence technology to convert vague conditions into specific search queries.

[1721] A "search query" is a keyword or phrase that a user enters into a search engine to retrieve specific information.

[1722] "E-commerce platform" refers to a website or application for buying and selling products and services over the Internet.

[1723] "Crawling" is the process of automatically visiting web pages on the Internet and collecting data.

[1724] "Product data" is a data set that contains detailed product information, such as price, description, and reviews.

[1725] "Filtering" is a process for narrowing down data based on specific conditions.

[1726] "Ranking" is the process of ranking data based on specific criteria.

[1727] "Personalized suggestions" analyze a user's past behavior and preferences to suggest products and services that are best suited to each individual user.

[1728] A "feedback loop" is a system that collects user reactions and behavior and reflects them in the next service provision.

[1729] The system for realizing this invention mainly includes a terminal that processes user input, a server that analyzes ambiguous conditions and generates optimal search queries, and a crawling function that aggregates product data from e-commerce platforms based on the generated queries.

[1730] User input via terminal

[1731] The user inputs their desired conditions and requests into the terminal. For example, the user might input "I want autographed merchandise of a famous athlete." Once the input is confirmed, the terminal sends this information to the server. The terminal can be a smartphone, tablet, PC, or other device.

[1732] Server-based analysis of ambiguous conditions

[1733] The server passes the vague conditions received from the device to a generative AI model. For example, the BART transformer model is used as the generative AI model. The model converts the vague conditions into a specific search query. For example, the condition "I want signed merchandise by a famous athlete" is converted to "famous athlete signed merchandise."

[1734] Crawling product data from e-commerce sites

[1735] The server uses the generated search query to crawl multiple e-commerce platforms (e.g., Amazon and Rakuten) to retrieve relevant product data, including information such as price, description, reviews, and ratings.

[1736] Filtering and Ranking Data

[1737] The acquired product data is filtered on the server and ranked based on user-specified criteria (price range, ratings, etc.) The filtering and ranking algorithms are implemented in Python, for example.

[1738] Providing optimal search results

[1739] The server sends the filtered and ranked product data to the device, which displays it to the user, who can then browse the best products in a list or grid view and select specific products.

[1740] Personalized product recommendations

[1741] The server analyzes the user's past purchasing behavior and search history and uses a generative AI model to make personalized suggestions, thereby providing users with more relevant and personalized product suggestions.

[1742] Formation of a feedback loop

[1743] The device sends user operation information and feedback to the server, which then analyzes the feedback and creates a feedback loop to improve the accuracy of the next search.

[1744] Specific examples

[1745] For example, if a user types in "latest smartphone case recommendations," the server converts this to "latest smartphone cases recommendations," retrieves smartphone case information from Amazon and Rakuten, and displays a list.

[1746] Prompt Sentence Examples

[1747] "Generate what query would be converted when a user types 'I want autographed merchandise of a famous athlete' into a product search app."

[1748] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1749] Step 1:

[1750] Get the conditions entered by the user

[1751] The device accepts input from the user. Specifically, the user enters vague keywords such as "I want autographed merchandise of a famous athlete" into the product search bar. Once this input is confirmed, the device converts the information into a data format such as JSON and sends the data to the server.

[1752] Input: Vague keywords entered by the user

[1753] Output: Keyword information converted into data format such as JSON

[1754] Step 2:

[1755] Convert vague conditions into specific search queries

[1756] The server passes the vague conditions received from the device to a generative AI model. The generative AI model (e.g., the BART transformer model) analyzes the conditions and converts them into specific search queries. For example, "I want autographed merchandise by a famous athlete" is converted to "famous athlete signed merchandise."

[1757] Input: Vague keywords in JSON format sent from the terminal

[1758] Output: A specific search query transformed by the generative AI model

[1759] Step 3:

[1760] Crawl multiple e-commerce platforms based on a search query

[1761] The server uses the generated search query to access the APIs of multiple e-commerce platforms (e.g., Amazon and Rakuten) and crawl product data, including product names, prices, descriptions, reviews, ratings, etc.

[1762] Input: A specific search query transformed by a generative AI model

[1763] Output: Product data obtained from multiple e-commerce platforms

[1764] Step 4:

[1765] Filter and rank your product data

[1766] The server aggregates the acquired product data and filters it based on the user's specified conditions (price range, ratings, etc.), and then ranks the filtered product data based on criteria such as ratings and price.

[1767] Input: Product data obtained from multiple e-commerce platforms

[1768] Output: Filtered and ranked product data

[1769] Step 5:

[1770] Displaying the most relevant product information to users

[1771] The server transmits the filtered and ranked product data to the terminal, which displays the data to the user in a format such as a list view or a grid view, allowing the user to review and select products.

[1772] Input: Filtered and ranked product data

[1773] Output: Product information displayed in list or grid view format

[1774] Step 6:

[1775] Make personalized product recommendations based on past purchasing behavior and search history

[1776] The server analyzes the user's past purchasing behavior and search history, and uses a generative AI model based on that information to make personalized product suggestions, allowing users to be recommended products that match their preferences.

[1777] Input: User's past purchasing behavior and search history

[1778] Output: personalized product recommendations

[1779] Step 7:

[1780] Forming a feedback loop

[1781] The device sends user operation information and feedback to the server, which analyzes this feedback and reflects it in order to improve the accuracy of the next search, allowing the user to obtain more accurate search results.

[1782] Input: User operation information and feedback

[1783] Output: Data reflected to improve search accuracy next time

[1784] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1785] In an embodiment of the present invention, a system is provided that allows users to easily search for desired products using vague search criteria and obtain optimal results. Furthermore, the system recognizes the user's emotions and improves the accuracy of search queries based on those emotions, thereby providing more personalized product offers. A specific embodiment of the system and its operation are described below.

[1786] Getting User Input

[1787] User

[1788] Users enter vague criteria into the search bar on their device (smartphone, tablet, PC, etc.). For example, they enter an abstract request such as "I want autographed merchandise of a famous athlete." Once they've finished entering the information, they click the "Search" button.

[1789] Terminal

[1790] The terminal receives the information entered by the user, converts the data into an appropriate data format such as JSON, and sends it to the server.

[1791] Condition analysis and emotion recognition

[1792] server

[1793] The server receives the ambiguous condition data sent from the device, passes this data to the generative AI model, and begins the analysis process.

[1794] Generative AI Models

[1795] The generative AI model analyzes the user's vague search criteria and converts them into a specific search query, which is then used to provide optimal search results.

[1796] Emotion Engine

[1797] The emotion engine analyzes the user's emotions at the time of input from voice, text input, facial expressions, etc. Based on the results of this analysis, the generative AI model optimizes its condition analysis. For example, if the user is excited, it will make more proactive product suggestions, while if the user is calm, it will emphasize more detailed information.

[1798] Crawling e-commerce sites

[1799] server

[1800] The server uses the generated search query and the sentiment data obtained from the sentiment engine to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.) and sends a search query to each platform to retrieve related product data.

[1801] Filtering and ranking results

[1802] server

[1803] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. This dataset is then filtered to narrow down the list based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on criteria such as price, popularity, and review ratings.

[1804] Providing optimal search results

[1805] server

[1806] The server transmits the filtered and ranked product data to the terminal.

[1807] Terminal

[1808] The device displays the received product data to the user in a list view, grid view, or other display format, allowing the user to easily find the best product from a wide range of options.

[1809] Personalized recommendations and feedback loops

[1810] server

[1811] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. Information about the user's selection of specific products and their details is stored and used to improve the accuracy of the next search criteria. User emotional data is also stored and reflected in future searches and suggestions.

[1812] user

[1813] Users can select a product from the product list displayed on their device and proceed to the details page or purchase page. They can also enter new search criteria and search again.

[1814] Specific examples

[1815] For example, suppose a user who wants to find autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete" and seems to be having fun. In this case, the system will perform the following actions.

[1816] 1. User Input

[1817] The user types "I want autographed merchandise from a famous athlete" into the device.

[1818] 2. Condition Analysis and Emotion Recognition

[1819] The server passes the input to a generative AI model for analysis, resulting in the search query "famous athlete signed merchandise."

[1820] The emotion engine analyzes the user's positive emotions and, based on the generated query, prioritizes products with more positive reviews.

[1821] 3. E-commerce site crawling

[1822] The server crawls multiple e-commerce platforms and collects product data.

[1823] 4. Filtering and Ranking Results

[1824] The server filters the collected product data and prioritizes and ranks products that have high ratings and many positive reviews that reflect the user's enjoyment emotions.

[1825] 5. Search result presentation

[1826] The server sends the filtered and ranked results to the device.

[1827] The device displays the appropriate product list to the user.

[1828] 6. Personalized Offers

[1829] The server analyzes the user's behavioral and emotional data to make more relevant personalized suggestions in the future.

[1830] In this way, a system is provided that allows users to easily find the products they are looking for without knowing the specific product names or detailed conditions, and furthermore, allows them to receive appropriate product suggestions based on their emotions.

[1831] The processing flow will be explained below.

[1832] Step 1:

[1833] User: Enters vague criteria such as "I want autographed merchandise of a famous athlete" into the device's search bar. Once entered, presses the "Search" button.

[1834] Step 2:

[1835] Terminal: Receives the search criteria entered by the user, converts the data into an appropriate format (e.g., JSON), and sends it to the server. An example of sending data is {"query":"I want autographed merchandise by a famous athlete"}.

[1836] Step 3:

[1837] Server: Receives search criteria data sent from the device and temporarily stores this data in memory.

[1838] Step 4:

[1839] Server: Passes the received data to the generative AI model and begins the analysis process. The generative AI model analyzes the user's vague search criteria and converts them into an appropriate search query (e.g., "famous athlete signed merchandise").

[1840] Step 5:

[1841] Emotion engine: Analyzes the emotions at the time of input from the user's voice, text input, facial expressions, etc. The emotion engine outputs the analysis results and optimizes the condition analysis of the generative AI model.

[1842] Step 6:

[1843] Server: Stores the search queries generated by the generative AI model and the sentiment analysis data from the sentiment engine, and prepares them to be sent to multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.).

[1844] Step 7:

[1845] Server: Uses generated search queries to access APIs and crawl the web for each e-commerce platform, and collects product data (price, description, ratings, etc.) from each platform.

[1846] Step 8:

[1847] Each e-commerce site: Returns a list of products based on the search query to the server through the provided API or web interface. For example, the results are returned in the format {"site": "Amazon", "results": [...]}.

[1848] Step 9:

[1849] Server: Aggregates product data collected from each e-commerce platform to form a large product dataset, which is temporarily stored in memory.

[1850] Step 10:

[1851] Server: Filters the aggregated product data. Based on the user's criteria (price range, ratings, etc.), it runs a process to narrow down the list to only products that match the criteria. It also prioritizes products based on sentiment data (e.g., if the user seems happy, prioritize products with many positive reviews).

[1852] Step 11:

[1853] Server: Ranks the filtered products. Ranking criteria include price, popularity, review rating, etc. Generates a ranked product list.

[1854] Step 12:

[1855] Server: Sends the filtered and ranked product list to the device. Example data sent includes {"ranked_products": [...]}.

[1856] Step 13:

[1857] Device: The received product list is displayed to the user in a format appropriate for the user interface, such as list view or grid view.

[1858] Step 14:

[1859] Users can select the product they are interested in from the list of products presented and proceed to the detail page or purchase page. If they do not like the product, they can enter new search criteria and search again.

[1860] Step 15:

[1861] Server: Records user behavior (selection, confirmation of details, purchase, etc.) and emotional data, and stores it as analytical data to improve search accuracy next time.

[1862] Step 16:

[1863] Server: Prepares to make future personalized product recommendations based on the analyzed user and sentiment data.

[1864] This series of steps allows users to easily find the products they are looking for and allows the system to continuously improve its accuracy.

[1865] Example 2

[1866] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1867] In conventional e-commerce platforms, when users search for products using vague criteria, they often encounter problems such as insufficient search results or being unable to find the desired product. Furthermore, because the platform does not take into account the user's purchasing motivation or emotions, personalized product suggestions are not possible, often resulting in low satisfaction. The present invention aims to solve these problems and provide users with optimal product search results and personalized suggestions.

[1868] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1869] In this invention, the server includes a means for analyzing vague search terms entered by a user based on a generative AI model and converting them into an appropriate search query, an emotion recognition means for analyzing the user's emotional state, and a means for crawling multiple e-commerce platforms and aggregating product data based on the converted search query. This allows a specific search query appropriate to the user's intent to be generated, even if the user enters vague search terms, and the aggregated product data can be ranked based on the user's specified search terms and emotional state. Furthermore, the system analyzes the user's past behavioral data and emotional data to provide personalized product suggestions, thereby increasing user satisfaction.

[1870] "Ambiguous search conditions entered by the user" refers to non-specific search conditions entered by the user through the terminal.

[1871] A "generative AI model" refers to an artificial intelligence model that analyzes vague conditions entered by users and converts them into appropriate search queries.

[1872] A "search query" refers to a specific search term or phrase that a generative AI model generates by analyzing ambiguous conditions.

[1873] "Emotion recognition means" refers to technology that analyzes the emotional state of a user from their input or interaction, including, for example, voice, text input, and facial expression recognition.

[1874] An "e-commerce platform" is a website or application that enables the buying and selling of goods and services online, such as Amazon or eBay.

[1875] "Crawling" refers to the act of a server automatically searching for information on a particular website and collecting data.

[1876] "Product Data" refers to all product-related information obtained from the e-commerce platform, such as product name, price, review rating, etc.

[1877] "Filtering" refers to the process of narrowing down collected product data based on user-specified conditions.

[1878] "Ranking" refers to the process of ranking filtered product data based on specific criteria, such as price, popularity, or review rating.

[1879] "Specified conditions" refer to conditions that a user sets to narrow down search results, such as a price range or rating range.

[1880] "Individualized product proposals" refer to product proposals that are optimized for each individual user based on the user's past behavioral and emotional data.

[1881] A "feedback loop" refers to a system that receives feedback from users and uses that information to improve the accuracy of subsequent searches and suggestions.

[1882] The present invention is a system for providing optimal search results when a user searches for products under vague conditions. Furthermore, the system recognizes the user's emotions and improves the accuracy of the search query based on the emotions, thereby providing more personalized product suggestions. A specific embodiment of the system and its operation are described below.

[1883] 1. Getting User Input

[1884] User

[1885] Users enter vague criteria into the search bar on their device (smartphone, tablet, PC, etc.), for example, an abstract request such as "I want autographed merchandise of a famous athlete." Once they're done, they click the "Search" button.

[1886] Terminal

[1887] The device acquires the conditions entered by the user, converts this input data into an appropriate data format such as JSON, and sends it to the server. At this time, when the user has completed entering the information on the device, the device sends an HTTP request to the server.

[1888] 2. Condition analysis and emotion recognition

[1889] server

[1890] The server receives vague condition data sent from the device. This data is passed to the generative AI model as a prompt. For example, the data {"query": "I want autographed merchandise of a famous athlete"} is sent.

[1891] Generative AI Models

[1892] The generative AI model analyzes the received prompt and converts the user's vague search criteria into a specific search query, such as "famous athlete signed merchandise." This converted query is cached on the server and can be reused.

[1893] Emotion Engine

[1894] The emotion engine analyzes user input (voice, text, facial expressions, etc.) to recognize the user's emotional state. For example, facial recognition technology can be used to determine whether a user is happy when typing. This emotional data influences the analysis of the generative AI model.

[1895] 3. Crawling e-commerce sites

[1896] server

[1897] The server uses the converted search query and sentiment data to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.). For each platform, it sends a new search query and collects related product data. For example, it uses the Amazon API to send the query "famous athlete signed merchandise" and retrieves the corresponding product data.

[1898] 4. Filtering and Ranking Results

[1899] server

[1900] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. This dataset is then filtered to narrow down the results based on user-specified criteria (price range, ratings, etc.). Specifically, it selects highly rated products within a certain price range. It also prioritizes products with proactive product suggestions and detailed information based on emotional data. For example, for users with a positive emotional state, it prioritizes products with many positive reviews.

[1901] 5. Providing optimal search results

[1902] server

[1903] The server sends the filtered and ranked product data to the terminal, for example, returning a product list in JSON format.

[1904] Terminal

[1905] The device displays the received product data to the user. The display format can be selected from list view, grid view, etc. For example, in list view, the product name, image, price, and review rating are displayed.

[1906] 6. Personalized recommendations and feedback loops

[1907] server

[1908] The server analyzes users' past purchasing behavior and search history to make personalized product suggestions. It also accumulates behavioral data, such as when users click on a specific product to view its details or make a purchase. This data is analyzed and saved along with emotional data to help improve the accuracy of the next search.

[1909] User

[1910] Users can select products they are interested in from the products displayed on their devices, proceed to the details and purchase process, or enter new search criteria and search again.

[1911] Specific examples

[1912] For example, suppose a user searching for autographed merchandise from a famous athlete enters the search criteria "I want autographed merchandise from a famous athlete" and seems to be having fun. In this case, the system performs the following specific actions:

[1913] 1. User Input

[1914] The user types "I want autographed merchandise from a famous athlete" into the device.

[1915] The user clicks the Search button.

[1916] 2. Condition Analysis and Emotion Recognition

[1917] The server passes the input data to the generative AI model.

[1918] A generative AI model generates the search query "famous athlete signed merchandise."

[1919] The emotion engine analyzes the user's happy facial expression.

[1920] 3. E-commerce site crawling

[1921] A server sends search queries to multiple e-commerce platforms to collect product data.

[1922] 4. Filtering and Ranking Results

[1923] The server filters the collected product data and prioritizes products with positive reviews and high ratings.

[1924] 5. Search result presentation

[1925] The server transmits the ranked product data to the terminal.

[1926] The device displays the product list to the user in a list view format.

[1927] 6. Personalized Offers

[1928] The server analyzes the user's behavioral and emotional data to make more appropriate personalized suggestions in the future.

[1929] Example prompt sentence:

[1930] "I want autographed merchandise from famous athletes."

[1931] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1932] Step 1:

[1933] Getting User Input

[1934] The user enters vague search criteria into the search bar of their device (smartphone, tablet, PC, etc.). For example, they might type "I want autographed merchandise of a famous athlete." Then, they click the "Search" button. The input data is in text format.

[1935] Step 2:

[1936] Sending data from the device to the server

[1937] The device converts the user-entered criteria into an appropriate data format, such as JSON. The device then sends this data to the server using an HTTP request, such as {"query": "I want autographed merchandise by a famous athlete"}. This data is provided by the user as input.

[1938] Step 3:

[1939] Condition Analysis

[1940] The server passes the vague condition data received from the device to the generative AI model. The generative AI model analyzes this data and converts it into a specific search query. For example, the vague condition "I want signed merchandise by a famous athlete" is converted into the specific search query "famous athlete signed merchandise." This converted query is output as search data.

[1941] Step 4:

[1942] Emotion Analysis

[1943] The server uses an emotion engine to analyze the user's input format (voice, text, facial expression, etc.). The emotion engine analyzes emotions from the user's facial expression and voice data at the time of input and returns the results to the server. For example, facial expression analysis can yield an emotion result of "looking happy." This emotion data becomes input for use in optimizing search queries.

[1944] Step 5:

[1945] Crawling e-commerce sites

[1946] The server uses the generated search query and the analyzed sentiment data to crawl multiple e-commerce platforms (e.g., Amazon, eBay, Rakuten, etc.). During the crawling, it uses the platform's API or scraping technology to send the search query and retrieve the corresponding product data. For example, it sends the search query "famous athlete signed merchandise" to Amazon to retrieve a list of related products. This product data is output as a result from the e-commerce site.

[1947] Step 6:

[1948] Filtering and ranking results

[1949] The server aggregates product data obtained from each e-commerce platform to form a large product dataset. Products are filtered from this dataset based on user-specified criteria (price range, ratings, etc.). The filtered product data is then ranked based on emotion data. For example, products with many positive reviews are prioritized for users with positive emotions. The filtered and ranked product data is output as an optimal product list.

[1950] Step 7:

[1951] Providing optimal search results

[1952] The server sends the filtered and ranked product data to the device, for example in the format {"products": [listed product data]}. This data is returned to the user as the best search results.

[1953] Terminal

[1954] The device displays the received product data to the user. The display format can be selected from list view, grid view, etc. For example, list view displays the product name, image, price, and review rating. This displayed product information is provided in a format that is easy for the user to view.

[1955] Step 8:

[1956] Personalized recommendations and feedback loops

[1957] The server analyzes the user's past purchasing behavior and search history to make personalized product suggestions. It also accumulates behavioral data, such as when a user clicks on a specific product to view its details page or make a purchase. This data is analyzed to help improve the accuracy of the next search criteria. For example, if a user repeatedly takes positive action regarding "signature goods," similar products will be included in the next suggestion. This feedback data is output as the user's behavioral pattern.

[1958] User

[1959] Users can select products they are interested in from the displayed products and proceed to the purchase process. They can also search again by entering new search criteria, improving the user experience. These new search criteria are then processed as data to be input again into the system.

[1960] As described above, the system of the present invention can efficiently carry out a series of processes from vague conditions to specific product proposals.

[1961] (Application example 2)

[1962] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1963] Existing search systems have difficulty finding appropriate search results for vague user inputs, making it difficult for users to smoothly find the products they are looking for. Furthermore, the lack of personalized suggestions based on user emotions and past behavior makes it difficult to provide a more personalized shopping experience. To solve these problems, there is a need for a search system that takes user emotions into account.

[1964] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving ambiguous conditions entered by a user and transmitting them to the server, means for the server to analyze the ambiguous conditions received based on a generative AI model and convert them into an appropriate search query, means for the server to crawl multiple e-commerce platforms and aggregate product data based on the converted search query, means for the server to filter the aggregated product data using an emotion recognition engine taking into account the user's emotional data and rank it based on the user's specified conditions, and means for the server to transmit the filtered and ranked product data to a terminal, which then displays it to the user. This enables users to accurately search for products even with ambiguous conditions and receive more personalized product suggestions based on their emotions.

[1965] "User" means an individual or organization that uses the System to search for and purchase products.

[1966] "Vague terms" are abstract search queries or requests entered by users without specifying specific products or specifications.

[1967] "Server" refers to a computer system that receives input data from users and performs processes such as analysis using generative AI models, emotion recognition engines, crawling e-commerce platforms, filtering and ranking data, etc.

[1968] A "generative AI model" is a type of artificial intelligence that analyzes ambiguous conditions and converts them into appropriate search queries, and is based on natural language processing.

[1969] An "emotion recognition engine" is a technology for analyzing emotions from non-verbal information such as a user's facial expressions and voice.

[1970] An "e-commerce platform" is an online marketplace for buying and selling goods and services over the Internet.

[1971] "Crawling" is the process of automatically gathering information on a website.

[1972] "Filtering" is the process of removing unnecessary data from collected product data based on specific criteria and extracting only the necessary data.

[1973] "Ranking" is the process of ranking the filtered product data based on specific evaluation criteria.

[1974] A "terminal" is a device that a user uses to access the system, including smartphones and computers.

[1975] A "feedback loop" is a cyclical process that analyzes information collected from users (e.g., search history, sentiment data) to improve the accuracy of future search criteria and suggestions.

[1976] The system for implementing this invention allows users to search for products under vague conditions and optimizes the results based on emotions. The specific configuration and operation of this system will be described below.

[1977] System Configuration

[1978] Hardware

[1979] Smartphone: A device with a built-in camera and microphone that allows users to access the system.

[1980] Server: A computer system equipped with a high-speed processor and capable of processing large amounts of data. Cloud-based operation is recommended.

[1981] software

[1982] Generative AI model: A natural language processing model to convert users' vague search terms into specific search queries. In this case, we use OpenAI's GPT-4.

[1983] Emotion recognition engine: An engine that analyzes the user's facial expressions and voice to recognize emotions. It uses Google Cloud Vision API and Microsoft Azure Face API.

[1984] E-commerce site crawling library: A library for collecting product data from e-commerce platforms. It uses BeautifulSoup (Python) and other tools.

[1985] Cloud platform: Infrastructure for operations and data processing. We recommend using Amazon Web Services (AWS).

[1986] System Operation

[1987] 1. Getting User Input

[1988] Users enter vague terms into the search bar on their smartphone and press the search button.

[1989] Example: "I want new shoes."

[1990] 2. Emotion recognition

[1991] The device's camera and microphone are used to collect the user's facial expressions and voice.

[1992] The emotion recognition engine analyzes this data and recognizes the user's emotions.

[1993] Example: A user looks happy.

[1994] 3. Condition analysis

[1995] Vague search queries are fed into a generative AI model, which converts them into specific search queries.

[1996] Example: "new shoes"

[1997] 4. Crawling e-commerce sites

[1998] The server uses the generated search query to crawl multiple e-commerce platforms to collect product data.

[1999] The collected data is converted into a unified format and aggregated on the server.

[2000] 5. Filtering and Ranking Results

[2001] The aggregated product data is filtered taking into account user sentiment data.

[2002] The filtered data is then ranked based on criteria such as price, popularity, and review ratings.

[2003] 6. Presentation of search results

[2004] The server transmits the filtered and ranked data to the terminal.

[2005] The terminal displays a list of products that are most suitable for the user.

[2006] 7. Creating feedback loops

[2007] User selections, search history, emotional data, etc. are stored on the server to improve search accuracy from the next time onwards.

[2008] Prompt Sentence Examples

[2009] Here is an example of a prompt for the generative AI model:

[2010] A user searched for: "I want new shoes." The user's sentiment seems to be happy. Convert this into a specific search query.

[2011] In this way, users can easily find the products they are looking for without having to enter specific product names or detailed search criteria. Furthermore, the system's emotion-based product suggestions enable a more personalized shopping experience.

[2012] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2013] Step 1: Getting User Input

[2014] The user enters vague search criteria into the device, for example, "I want new shoes." Once the search is complete, the user presses the search button. The device converts the input data into JSON format and sends it to the server.

[2015] Input: User's vague search criteria

[2016] Output: Input data in JSON format

[2017] Specific actions: Enter search criteria in the search bar and press the button.

[2018] Step 2: Emotion Recognition

[2019] The device uses a built-in camera and microphone to collect the user's facial expressions and voice. This data is sent to an emotion recognition engine to analyze the user's emotions. The analysis results are sent to the server in JSON format.

[2020] Input: User's facial expression and voice data

[2021] Output: Emotion data in JSON format

[2022] Specific actions: Capture facial expressions and voice using a camera and microphone.

[2023] Step 3: Condition analysis

[2024] The server receives the search criteria and emotion data sent from the device, passes this data to the generative AI model, and converts it into a specific search query. The generative AI model analyzes the prompt and generates an appropriate search query.

[2025] Input: Fuzzy search terms and sentiment data in JSON format

[2026] Output: Specific search query

[2027] Specific operation: The generative AI model analyzes the prompt sentence and generates a search query.

[2028] Step 4: Crawl your e-commerce site

[2029] The server crawls multiple e-commerce platforms using the generated specific search query, collects product data using libraries such as BeautifulSoup, converts the collected data into a unified format, and aggregates it on the server.

[2030] Input: A specific search query

[2031] Output: Aggregated product data in a unified format

[2032] Specific operation: Collects product data using e-commerce site crawl libraries.

[2033] Step 5: Filter and rank the results

[2034] The server receives the aggregated product data and filters it using an emotion recognition engine, taking into account the user's emotional data. The filtered data is then ranked based on criteria such as price, ratings, and reviews.

[2035] Input: Aggregated product data and sentiment data in a unified format

[2036] Output: Filtered and ranked product data

[2037] Specific behavior: Analyze product data and optimize based on user sentiment.

[2038] Step 6: Presenting search results

[2039] The server sends the filtered and ranked product data to the device, which then visually displays this data to the user in a format such as a list view or grid view.

[2040] Input: Filtered and ranked product data

[2041] Output: The optimal product list shown to the user

[2042] Specific operation: Display the product list on the terminal screen.

[2043] Step 7: Creating a feedback loop

[2044] The server stores user behavioral history and emotional data to improve the accuracy of search criteria and suggestions from the next time onward, allowing users to receive more personalized search results.

[2045] Input: User behavior history, emotional data

[2046] Output: Feedback to improve accuracy using accumulated data

[2047] Specific actions: Accumulate behavioral history and emotional data and reflect them in the next proposal.

[2048] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[2049] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[2050] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[2051] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[2052] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[2053] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[2054] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[2055] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[2056] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[2057] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[2058] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[2059] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[2060] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[2061] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[2062] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[2063] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[2064] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[2065] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[2066] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[2067] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to t...

Claims

1. means for receiving vague conditions entered by a user and transmitting them to a server; A means for the server to analyze the ambiguous conditions received based on a generative AI model and convert them into appropriate search queries; a server crawling a plurality of e-commerce platforms and aggregating product data based on the converted search queries; a means for the server to filter the aggregated product data and rank them based on user-specified criteria; The system includes a means for the server to transmit the filtered and ranked product data to the terminal, and for the terminal to display the data to the user.

2. 2. The system according to claim 1, wherein the server includes means for analyzing the user's past purchasing behavior and search history and making personalized product suggestions.

3. 2. The system according to claim 1, further comprising means for the terminal to send feedback from the user to the server, forming a feedback loop for improving the accuracy of the next search.

Images