System

Abstract

A system is provided.SOLUTION: A system comprising: means for collecting merchandise reviews based on a query received from a user; means for analyzing the collected merchandise reviews using natural-language processing techniques to classify the reviews and extract key phrases; means for evaluating the reliability of the reviews and filtering out stealth marketing and false reviews; means for generating an easy-to-read summary from the analyzed reviews; means for generating a personalized review based on the user's browsing and purchasing histories; and means for displaying the generated summary on the user's device.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] The e-commerce market currently has a huge number of product reviews, and consumers spend a lot of time and effort making purchasing decisions. Stealth marketing and fake reviews are also common, making it difficult for consumers to discern reliable information. This situation is damaging consumers' purchasing experience and negatively impacting the credibility of e-commerce platforms. Therefore, there is a need for platforms that allow consumers to select products quickly and easily and provide reliable review information. [Means for solving the problem]

[0005] The present invention solves the above-mentioned problems by providing a system including: means for collecting product reviews based on search queries received from users; means for analyzing the collected product reviews using natural language processing technology, categorizing the reviews, and extracting key phrases; means for evaluating the reliability of the reviews and filtering out stealth marketing and false reviews; means for generating easy-to-read summaries from the analyzed reviews; means for generating personalized reviews based on the user's browsing history and purchase history; and means for displaying the generated summaries on the user's device. This allows consumers to select products based on reliable review information in a short amount of time, thereby improving the purchasing experience.

[0006] A "search query" is a keyword or phrase related to a product or category that a user is interested in, and is the input information used to search for product reviews based on this.

[0007] "Natural language processing technology" is a technology for analyzing text data and understanding meaning and emotions, and is used to classify reviews, extract key phrases, and analyze sentiment.

[0008] "Classification" is the process of dividing collected product reviews into categories such as positive, negative, and neutral.

[0009] A "key phrase" is a phrase that indicates particularly important information or evaluation points in a review.

[0010] "Sentiment analysis" is a technique that analyzes the emotional tone contained in reviews and identifies positive, negative, or neutral emotional states.

[0011] "Trustworthiness scoring" is the process of evaluating and quantifying whether a review is trustworthy.

[0012] "Filtering" is the process of eliminating unnatural or false reviews based on the analysis results.

[0013] A "summary" is information that extracts and succinctly summarizes the main opinions and evaluation points from a large number of reviews.

[0014] "Personalized reviews" are review information that is optimized for each individual user based on the user's browsing history and purchasing history.

[0015] A "terminal" is a device used by a user to view information, and includes smartphones, personal computers, etc. [Brief explanation of the drawings]

[0016] [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

[0017] 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.

[0018] First, the terms used in the following description will be explained.

[0019] 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).

[0020] 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.

[0021] 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.

[0022] 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.

[0023] 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."

[0024] [First embodiment]

[0025] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0026] 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.

[0027] 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).

[0028] 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.

[0029] 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.

[0030] 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.

[0031] 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.

[0032] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0033] 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.

[0034] 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.

[0035] 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.

[0036] 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."

[0037] The present invention provides a system that allows users to quickly and easily select product reviews on an e-commerce site. This system is implemented using the following programs in particular.

[0038] Program processing explanation

[0039] 1. User searches for product reviews

[0040] A user visits an e-commerce site and enters keywords related to a specific product name or category into the search bar. Based on this search query, the server begins searching for relevant reviews. For example, a user enters "smartphone case" and clicks the search button.

[0041] 2. The server collects reviews

[0042] The server collects relevant product reviews from major e-commerce sites based on the search query entered by the user. The reviews are retrieved via API and stored in an internal database. This results in the collection of hundreds of reviews related to "smartphone cases."

[0043] 3. Review Analysis Using Natural Language Processing (NLP)

[0044] The server analyzes the collected reviews using natural language processing technology, which performs the following processes:

[0045] Review categorization: Categorize reviews as positive, negative, or neutral. For example, based on the content of the review, the part that says "high durability" can be categorized as positive, and conversely, the part that says "high price" can be categorized as negative.

[0046] Key phrase extraction: Extracting important information from each review, such as "nice design" or "durable and durable."

[0047] Sentiment analysis: Analyzing the tone of a review to determine emotional intensity, for example, distinguishing between "very satisfied" and "somewhat dissatisfied" reviews.

[0048] 4. Filtering and Eliminating Stealth Marketing

[0049] The server then uses the analysis results to assess trustworthiness and filter out fraudulent reviews, including:

[0050] Unnatural pattern detection: Automatically detects reviews posted in large numbers from the same IP address or reviews with abnormally high ratings.

[0051] Trustworthiness scoring: Each review is assigned a certain score to filter out fraudulent reviews. For example, reviews submitted in large numbers in a short period of time are given a low score and are subject to elimination.

[0052] 5. Generating a Summary

[0053] The server generates an easy-to-read summary based on the filtered reviews. This allows users to quickly obtain only the main points of evaluation without having to read lengthy reviews. For example, it generates a summary such as, "This smartphone case is highly durable and has a well-received design, but is a little pricey."

[0054] 6. Providing personalized reviews

[0055] The server generates personalized reviews based on the user's browsing and purchasing history. For example, if a user has previously purchased products that emphasize durability, the server provides a summary that highlights information such as, "This case is particularly durable."

[0056] 7. Display summary on terminal

[0057] The generated summary is displayed on the user's device. Summary information about the product is displayed on the screen of a smartphone or PC, allowing the user to quickly check important information. For example, a message may appear on a smartphone screen saying, "Many users rate this case highly for its durability and design, and although the price is somewhat high, satisfaction is high."

[0058] Specific examples

[0059] When a user searches for "smartphone case" on their smartphone, the server collects relevant reviews from e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries. These summaries are then displayed as personalized summaries on the user's device, allowing the user to quickly and efficiently evaluate the product.

[0060] A system based on this claim would speed up understanding of product reviews, increase their reliability, and improve the consumer buying experience.

[0061] The processing flow will be explained below.

[0062] Step 1:

[0063] A user types keywords related to a specific product name or category into the search bar of an e-commerce site, and is ready to find reviews that match their interests.

[0064] Step 2:

[0065] The server collects relevant product reviews from major e-commerce sites based on the search query received from the user. The server uses an API to retrieve reviews that match the keywords and stores them in an internal database.

[0066] Step 3:

[0067] The server starts analyzing the collected product reviews by applying natural language processing (NLP) techniques. The server performs the following processes:

[0068] Classification of reviews: Positive, negative, or neutral.

[0069] Key phrase extraction: Extract important information and evaluation points.

[0070] Sentiment analysis: Analyze the emotional tone of each review.

[0071] Step 4:

[0072] The server uses the analysis results to assess the trustworthiness of reviews and filter out stealth marketing and fake reviews. This includes:

[0073] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with only unusually high ratings.

[0074] Trustworthiness Scoring: Each review is assigned a trustworthiness score to help weed out fraudulent reviews.

[0075] Step 5:

[0076] The server generates easy-to-read summaries from reviews that have passed the credibility assessment. The server concisely extracts key opinions and evaluation points and summarizes them in a format that users can quickly understand.

[0077] Step 6:

[0078] The server generates personalized reviews based on the user's browsing and purchasing history, analyzing the user's past preferences and reactions and providing individually optimized review information.

[0079] Step 7:

[0080] The terminal displays the generated summary information to the user, who can then visually view the generated summary on the screen of their smartphone or computer, allowing them to quickly check the product evaluation.

[0081] This is the specific processing flow of the "AI Review Summarizer" system, which enables users to quickly organize and accurately understand vast amounts of review information.

[0082] Example 1

[0083] 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."

[0084] When purchasing a product on an online e-commerce site, users need to find useful information from numerous reviews, but this task requires time and effort. Furthermore, some reviews may be fraudulent or unreliable, which can lead to misjudgments. Furthermore, because different users have different priorities, it is not always easy to find the most suitable review for each user. Therefore, there is a need for a fast and reliable review selection system that can provide personalized reviews tailored to users' preferences.

[0085] 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.

[0086] In this invention, the server includes means for collecting product reviews based on a search query received from a user, means for analyzing the collected product reviews using natural language processing technology, categorizing the reviews, and extracting key phrases, means for evaluating the reliability of the reviews and filtering out fraudulent reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized summaries based on the user's browsing history and purchase history, and means for providing the generated summaries to a display device. This allows users to quickly and efficiently obtain reliable product reviews and make purchasing decisions based on information that suits their needs.

[0087] A "search query" is a keyword or phrase that a user enters to search for information related to a particular product or category.

[0088] A "product review" is a written description of the product's usage and evaluation by a user who has purchased the product on an e-commerce site or the like.

[0089] "Natural language processing technology" is a technology that enables computers to understand and analyze human language and extract its meaning.

[0090] "Review categorization" is the process of dividing collected product reviews into categories based on their evaluation, such as positive, negative, or neutral.

[0091] A "key phrase" is a major phrase in a piece of text that expresses important information or a particular theme.

[0092] "Review credibility assessment" is the process of determining how accurate and trustworthy the collected reviews are.

[0093] "Fraudulent reviews" are false or exaggerated reviews posted with the intent to manipulate product sales.

[0094] "Filtering" is the process of removing unnecessary or fraudulent information from collected data.

[0095] A "summary" is a concise summary of the main points extracted from a lengthy review.

[0096] A "personalized summary" is a summary that is individually customized based on a particular user's preferences and purchasing history.

[0097] A "display device" is a device for visually presenting the generated summary or other information to a user, including a smartphone or personal computer.

[0098] This invention provides an information processing system that allows users to quickly and easily select product reviews on e-commerce sites. The system collects relevant reviews based on a user's search query, analyzes them using natural language processing (NLP), evaluates their credibility, generates summaries, and provides personalized summaries to users.

[0099] System Configuration

[0100] The system consists of the following main components:

[0101] 1. User device: The device where a user enters a search query and sees the results. This can be a smartphone or a computer.

[0102] 2. Server: A central processing unit that receives search queries from users and collects, analyzes, filters, and summarizes reviews. The server is made up of hardware with a high-performance processor and large memory capacity.

[0103] 3. Natural Language Processing Engine: Software for analyzing reviews. NLP tools such as Python's NLTK (Natural Language Toolkit) and Gensim are used.

[0104] Data processing and calculation

[0105] The server does the following:

[0106] 1. Review collection: The server collects relevant product reviews using the APIs of major e-commerce sites based on the search query received from the user. For example, if a user searches for "smartphone case," the server retrieves data via the APIs of Amazon and Rakuten Ichiba.

[0107] 2. Review Analysis: We analyze the collected reviews using natural language processing (NLP). We classify reviews as "positive," "negative," or "neutral," extract key phrases, and perform sentiment analysis. We use the NLTK and Gensim libraries to perform this analysis.

[0108] 3. Review credibility evaluation and filtering: The server evaluates the credibility of reviews based on the analysis results and filters out fraudulent reviews. The algorithms used include unnatural pattern detection and credibility scoring.

[0109] 4. Summary Generation: The server generates summaries from the filtered reviews. It also creates personalized summaries that emphasize different points for each user. It uses the Gensim library to generate the summary text.

[0110] Specific examples

[0111] For example:

[0112] When a user searches for "smartphone case" on their smartphone, the following prompt is entered into the system:

[0113] "I want to generate summaries from reviews about smartphone cases, classify them as positive or negative, and extract key phrases."

[0114] The server receives the search query and collects review information via APIs from Amazon and Rakuten. It then analyzes the reviews using NLTK and Gensim and classifies them as positive, negative, or neutral. It then scores the reviews' trustworthiness, filters out unnatural patterns, and generates a personalized summary tailored to the user's preferences. This summary is then displayed on the browser screen of a smartphone or PC, allowing users to easily check important information about the review.

[0115] This invention enables users to quickly and efficiently obtain highly reliable product reviews and make purchasing decisions based on information that meets their needs.

[0116] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0117] Step 1:

[0118] User searches for product reviews

[0119] A user accesses an e-commerce site using a smartphone or PC and enters keywords related to a specific product name or category into the search bar. For example, if the user enters "smartphone case," the search query is sent to the server when the user clicks the search button.

[0120] Input: The search query entered by the user (e.g., "smartphone case")

[0121] Output: The search query sent to the server

[0122] Step 2:

[0123] Server collects reviews

[0124] The server uses the APIs of major e-commerce sites to collect relevant product reviews based on the search query received from the user. Secure data acquisition is performed using an API key, and the acquired review information is stored in an internal database. For example, reviews on "smartphone cases" are collected using the APIs of Amazon and Rakuten Ichiba.

[0125] Input: The search query received by the server

[0126] Output: A database of collected review information

[0127] Step 3:

[0128] Review analysis using natural language processing (NLP)

[0129] The server analyzes the collected reviews using natural language processing (NLP) technology. Specifically, it performs the following processes:

[0130] Review Classification: Classify review text as positive, negative, or neutral. For example, "The case is very sturdy" is classified as positive, and "The price is too high" is classified as negative.

[0131] Keyphrase extraction: Important keyphrases are extracted using TF-IDF and Word2Vec algorithms. Examples include "good design" and "durable."

[0132] Sentiment analysis: Analyzes the sentiment of the review and evaluates the strength of the sentiment. For example, a "Very Satisfied" review will have a high positive score, while a "Slightly Dissatisfied" review will have a negative score.

[0133] Input: Collected review information

[0134] Output: Analyzed review classification, key phrases, and sentiment analysis data

[0135] Step 4:

[0136] Review credibility rating and filtering

[0137] The server evaluates the reliability of reviews based on the analysis results and performs filtering. Specifically, it performs the following processes:

[0138] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with only unusually high ratings.

[0139] Trustworthiness Scoring: Assign a score to each review and automatically filter out fraudulent reviews with low scores.

[0140] Input: Parsed review classification, key phrases, and sentiment analysis data

[0141] Output: Filtered data containing only reliable reviews

[0142] Step 5:

[0143] Generate a summary

[0144] The server generates summaries based on the filtered reviews. Using the Gensim library, it extracts key points from the review text and summarizes them in a concise format to create easy-to-read summaries. For example, it generates a summary such as, "This smartphone case is durable and has a good design, but is a little pricey."

[0145] Input: Filtered reviews

[0146] Output: Generated summary

[0147] Step 6:

[0148] Generate personalized reviews

[0149] The server generates a personalized summary based on the user's browsing and purchasing history. For a user who has previously purchased products that emphasize durability, the server generates a summary that emphasizes the fact that "this case is particularly durable."

[0150] Input: User browsing history, purchase history, filtered reviews

[0151] Output: Personalized Summary

[0152] Step 7:

[0153] Display summary to terminal

[0154] The generated summary is displayed on the user's device. Summary information about the product is displayed on the screen of a smartphone or computer, allowing the user to quickly check important information. Specifically, the smartphone browser or dedicated app screen displays the message, "Many users rate this case highly for its durability and design, and although the price is a little high, satisfaction is high."

[0155] Input: Generated summary, personalized summary

[0156] Output: Summary information displayed on the user's terminal

[0157] (Application example 1)

[0158] 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."

[0159] Online e-commerce sites contain a huge number of product reviews, making it difficult for users to quickly access the information they need. Conventional systems lack the ability to evaluate and filter the reliability of reviews, and many contain stealth marketing and fraudulent reviews. Furthermore, personalized reviews tailored to each user's needs are not provided, making it difficult for users to effectively utilize reviews. This can lead to consumers making incorrect decisions or wasting unnecessary time when choosing products.

[0160] 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.

[0161] In this invention, the server includes means for collecting product reviews based on a search query received from a user, means for analyzing the collected product reviews using natural language processing technology to categorize the reviews and extract key phrases, means for evaluating the reliability of the reviews and filtering out stealth marketing and false reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized reviews based on the user's browsing history and purchasing history, means for displaying the generated summaries on the user's device, means for analyzing the content of the reviews and categorizing them as positive, negative, or neutral, and means for analyzing detailed information about the products viewed by the user and generating prompt sentences. This allows users to quickly access reliable review summaries and efficiently obtain information tailored to their purchasing behavior.

[0162] A "search query" is a keyword or phrase that a user enters when searching for a product or information.

[0163] A "product review" is a record of the evaluation and impressions of a product made by a user who has purchased the product.

[0164] "Natural language processing technology" is a technology that allows computers to understand, analyze, and generate human language.

[0165] "Keyphrases" refer to important words or phrases within a review or document.

[0166] "Stealth marketing" refers to advertising activities that are carried out without consumers realizing that they are advertising.

[0167] A "false review" is a review that contains false information that differs from the actual performance or usability of the product.

[0168] A "summary" is a short summary that extracts the main points from a long document or review.

[0169] A "personalized review" is a review that is customized based on the interests and preferences of an individual user.

[0170] "Evaluating credibility" refers to the act of judging and evaluating the accuracy and reliability of collected reviews.

[0171] This invention provides a system that allows users to quickly and easily select product reviews on an e-commerce site. This system is realized through the following steps.

[0172] First, a user accesses an e-commerce site from a smartphone or computer and enters a search query related to a specific product name or category. This search query is sent to a server, which then collects relevant product reviews from the e-commerce site. The collected reviews are then stored in the server's database.

[0173] The server then analyzes the collected reviews using natural language processing (NLP) technology. This analysis includes categorizing reviews into positive, negative, and neutral categories and extracting key phrases. Specific analysis techniques used include tools such as the Python nltk library and scikit-learn.

[0174] The server then evaluates the credibility of reviews based on the analysis results and filters out fraudulent reviews. This evaluation is based on review posting frequency and IP address pattern analysis. It also applies algorithms to automatically detect stealth marketing and fake reviews. This ensures users have access to reliable review information.

[0175] The server generates summaries of the analyzed and filtered reviews. These summaries are short summaries that extract the main points of evaluation from the long reviews. For example, a summary might be generated that reads, "This smartphone case is durable and has a good design, but is a little pricey."

[0176] The server also generates personalized reviews based on the user's browsing and purchasing history. These personalized reviews provide information customized to the user's preferences and interests. For example, for a user who has previously purchased a durable product, the server may emphasize that "this case is particularly durable."

[0177] Finally, the generated summary and personalized review are displayed on the user's device. Users can view the concise review information on their smartphone or computer screen, allowing them to quickly and efficiently understand product ratings and obtain information that will help them make a purchase.

[0178] For example, when a user searches for "smartphone case" on their smartphone, the server collects relevant reviews from e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates a summary, which is then displayed as a personalized summary on the user's device.

[0179] An example prompt is, "Based on the product name searched by the user, collect relevant reviews, analyze the reviews using natural language processing, and generate summaries. In particular, display them separately from the perspectives of positive, negative, and neutral."

[0180] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0181] Step 1:

[0182] A user accesses an e-commerce site using a smartphone or computer and enters a search query related to a specific product name or category. The search query is sent from the device to the server. The input data at this time is the keywords or phrases entered by the user.

[0183] Step 2:

[0184] The server collects relevant product reviews from major e-commerce sites based on the received search query. The collection is done using an API, and the acquired review information is stored in the server's database. The output data at this time is raw review information.

[0185] Step 3:

[0186] The server uses natural language processing (NLP) technology to analyze the collected product reviews. Specifically, it categorizes the reviews into positive, negative, and neutral categories and extracts important key phrases. The NLP technology used includes Python's nltk library and scikit-learn. The input data is raw review information, and the output data is analyzed review information.

[0187] Step 4:

[0188] The server evaluates the reliability of the reviews based on the analysis results and filters out fraudulent reviews. The filtering method uses an algorithm that automatically detects large numbers of posts from the same IP address and reviews with abnormally high ratings. The input data is the analyzed review information, and the output data is the reliability-evaluated review information.

[0189] Step 5:

[0190] The server generates easy-to-read summaries from the reviews that have been evaluated for trustworthiness. The summary generation uses technology to extract the main evaluation points that have been analyzed and summarize them in a short form. For example, a summary such as "It is highly durable and has a popular design, but it is expensive" may be generated. The input data at this time is the review information that has been evaluated for trustworthiness, and the output data is the summary information.

[0191] Step 6:

[0192] The server generates personalized reviews based on the user's browsing history and purchase history. This utilizes information such as the user's past product preferences. For example, for a user who prioritizes durability, the server emphasizes information such as "This case is particularly durable." The input data at this time are summary information and user history information, and the output data is personalized summary information.

[0193] Step 7:

[0194] The generated summary and personalized review are displayed on the user's device. The user can view the concisely summarized review information on the screen of a smartphone or computer. The input data is the personalized summary information, and the output data is the summary information displayed on the user's device.

[0195] Step 8:

[0196] The server analyzes detailed information about the product viewed by the user and generates a prompt sentence. This prompt sentence is used to complement the user's next search query and review confirmation. For example, a prompt sentence may be generated that reads, "Based on the product name searched by the user, please collect related reviews, analyze the reviews using natural language processing, and generate summaries. Please display them in particular from the perspectives of positive, negative, and neutral." The input data at this time are browsing history information and review information, and the output data is the prompt sentence.

[0197] 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.

[0198] The present invention is a system that enables users to quickly and easily select product reviews on e-commerce sites, and also has the ability to recognize users' emotions and customize the review display. The specific program processing will be explained in detail below, with examples.

[0199] Program processing explanation

[0200] 1. User searches for product reviews

[0201] A user visits an e-commerce site and enters keywords related to a specific product name or category into the search bar, which prepares the site to retrieve reviews tailored to the user's interests. For example, a user enters "smartphone cases" and clicks the search button.

[0202] 2. The server collects reviews

[0203] The server collects relevant product reviews from major e-commerce sites based on the search query received from the user. The reviews are retrieved via API and stored in an internal database. For example, hundreds of reviews on "smartphone cases" are collected.

[0204] 3. Review Analysis Using Natural Language Processing (NLP)

[0205] The server analyzes the collected product reviews using natural language processing (NLP) techniques. It performs the following processes:

[0206] Review categorization: Categorize reviews as positive, negative, or neutral. For example, categorize reviews that say "high durability" as positive and reviews that say "high price" as negative.

[0207] Key phrase extraction: Extract important information and evaluation points, such as "good design" or "durable and durable."

[0208] Sentiment analysis: Analyzing the emotional tone of each review, for example, distinguishing between "Very Satisfied" and "Somewhat Dissatisfied."

[0209] 4. Filtering and Eliminating Stealth Marketing

[0210] The server evaluates the credibility of the reviews based on the analysis results and filters out stealth marketing and fake reviews. This includes the following actions:

[0211] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with an unusually high number of positive ratings.

[0212] Credibility scoring: Each review is assigned a credibility score to filter out fraudulent reviews. For example, reviews posted in large numbers in a short period of time will be assigned a low score and will be subject to elimination.

[0213] 5. Generating a Summary

[0214] The server generates easy-to-read summaries from reviews that pass the trust assessment. Users can quickly obtain only the main points of evaluation without having to read lengthy reviews. For example, a summary such as "This smartphone case is highly durable and has a well-received design, but is a little pricey" can be generated.

[0215] 6. Providing personalized reviews

[0216] The server generates personalized reviews based on the user's browsing and purchasing history. It analyzes the user's past preferences and reactions and provides individually optimized review information. In particular, for users who have previously preferred products that emphasize durability, it provides a summary such as "This case is particularly durable."

[0217] 7. Emotion Recognition with Emotion Engine

[0218] The server uses an emotion engine to recognize the user's emotion. It performs the following operations:

[0219] Facial recognition technology: Analyzes the user's facial expressions to identify their current emotional state. For example, it can detect when a user is unhappy.

[0220] Voice analysis: Analyzing the user's vocal utterances to identify their emotional state, for example, by analyzing the user's tone of voice and word choice.

[0221] 8. Customize review display based on sentiment

[0222] The server customizes the summary based on the information obtained from the emotion engine, providing review information that matches the user's emotional state. For example, if the user is in a positive mood, the server highlights the same positive reviews.

[0223] 9. Display summary on terminal

[0224] The generated summary is displayed on the user's device. The device visually presents the generated summary on the screen of a smartphone or PC. For example, a smartphone screen might say, "Many users rate this case highly for its durability and design, and although the price is a little high, satisfaction is high."

[0225] Specific examples

[0226] When a user searches for "smartphone case," the server collects related reviews and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries from reliable reviews. It then creates personalized reviews based on the user's browsing history and emotional state and displays them on the device. This allows users to quickly and accurately evaluate products.

[0227] The system according to this claim enhances the consumer buying experience by speeding up the understanding of product reviews and customizing the display according to the user's sentiment.

[0228] The processing flow will be explained below.

[0229] Step 1:

[0230] A user visits an e-commerce site and types keywords related to a specific product name or category into the search bar, generating a search query to find reviews that match the user's interests.

[0231] Step 2:

[0232] The server collects related product reviews from major e-commerce sites based on the search query received from the user. The server uses an API to retrieve reviews that match the keywords and stores them in an internal database. For example, it collects reviews related to "smartphone cases."

[0233] Step 3:

[0234] The server applies natural language processing (NLP) techniques to the collected product reviews to analyze them, including:

[0235] Review classification: Classify collected reviews as positive, negative, or neutral. For example, "The case is durable" is classified as positive, and "The price is high" is classified as negative.

[0236] Key phrase extraction: Extract key information and evaluation points from each review, such as "good design" or "durable and durable."

[0237] Sentiment analysis: Analyzing the emotional tone of a review to determine its strength, for example, distinguishing between "Very Satisfied" and "Slightly Dissatisfied."

[0238] Step 4:

[0239] The server evaluates the credibility of the reviews based on the analysis results and filters out stealth marketing and fake reviews. It does the following:

[0240] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with abnormally high ratings.

[0241] Trustworthiness Scoring: Each review is assigned a trustworthiness score to filter out fraudulent reviews. For example, reviews posted in large numbers in a short period of time are filtered out as low scores.

[0242] Step 5:

[0243] The server generates easy-to-read summaries from reviews that have passed the reliability assessment. The server succinctly summarizes key opinions and evaluation points so that users can understand them quickly. For example, it generates a summary such as, "This smartphone case is highly durable and has a well-received design, but the price is a little high."

[0244] Step 6:

[0245] The server generates personalized reviews based on the user's browsing and purchasing history. This analyzes the user's past preferences and reactions and provides optimized review information. For example, a user who prioritizes durability might be provided with a summary such as "This case is particularly durable."

[0246] Step 7:

[0247] The server uses an emotion engine to recognize the user's emotions, which involves the following steps:

[0248] Facial recognition technology: Analyzes the user's facial expressions to identify their current emotional state, for example, detecting dissatisfied facial expressions.

[0249] Voice analysis: Analyzing a user's voice to identify their emotional state, for example, determining their emotions based on their tone of voice and the way they speak.

[0250] Step 8:

[0251] The server customizes summaries based on the emotional state obtained from the emotion engine, providing users with review information that matches their emotional state. For example, if the user is in a positive mood, positive reviews will be highlighted.

[0252] Step 9:

[0253] The device displays the generated summary to the user. It is presented visually on the screen of a smartphone or computer. For example, it may display information such as, "Many users rate this case highly for its durability and design, and although it is somewhat expensive, satisfaction is high."

[0254] This is the specific processing flow of the "AI Review Summarizer" system, which allows users to quickly understand reviews and obtain information customized to their sentiment.

[0255] Example 2

[0256] 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."

[0257] Conventional product review management systems on e-commerce sites make it difficult for users to efficiently find useful information from the large number of reviews. Furthermore, stealth marketing and false reviews can make it difficult for users to make accurate decisions. Furthermore, the lack of functionality to customize reviews based on users' browsing history and emotional state makes it difficult to improve the user's purchasing experience.

[0258] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes: means for collecting product reviews based on a search query received from a user; means for analyzing the collected product reviews using natural language processing technology, categorizing the reviews, and extracting key phrases; means for evaluating the credibility of the reviews and filtering out fraudulent reviews; means for generating easy-to-read summaries from the analyzed reviews; means for generating personalized reviews based on the user's browsing history and purchasing history; means including an emotion engine for recognizing the user's emotions; means for customizing review display based on the emotion recognition results; and means for displaying the generated summaries on the user's device. This allows users to quickly and accurately obtain useful review information and make accurate decisions without being influenced by stealth marketing or fake reviews. Furthermore, the review display is customized based on the user's individual browsing history and emotional state, improving the purchasing experience.

[0259] A "search query" is a word or phrase that a user enters to search for specific information.

[0260] "Product reviews" are evaluations and comments written by users who have purchased a product about the quality and performance of that product.

[0261] "Natural language processing technology" is a technology that allows programs to understand, interpret, and generate human language.

[0262] "Classification" is the act of grouping collected data based on specific criteria.

[0263] "Keyphrase extraction" is the process of extracting important information and evaluation points from a text.

[0264] A "trustworthiness rating" is a rating used to determine whether the content of a review is trustworthy.

[0265] "Filtering" is the process of removing unwanted or harmful information.

[0266] "Summary generation" is the act of shortening a long piece of text and extracting and summarizing only the main points.

[0267] "Browsing history" is a record of pages and product information that a user has previously viewed on a website or app.

[0268] "Purchase history" is a record of products purchased by a user in the past.

[0269] A "personalized review" is a product review that is customized based on a user's individual tastes and preferences.

[0270] An "emotion engine" is a technology or system for recognizing and analyzing a user's emotional state.

[0271] "Customization" refers to adjustments or changes made to suit the needs and preferences of individual users.

[0272] A "terminal" is an electronic device used by a user to display information.

[0273] The present invention provides a system that enables users to quickly and easily select product reviews on an e-commerce site, and also has a function for recognizing user emotions and customizing the review display. A specific embodiment of this system will be described below.

[0274] 1. System Overview

[0275] The system performs a series of processes that allow users to search for product reviews and display them in a customized format. The system is designed on the premise that it is operated by a server, a user's terminal, and the user.

[0276] 2. Hardware and Software Configuration

[0277] Hardware

[0278] Server: A group of computers with high-performance processing power and large storage capacity. They are responsible for collecting and analyzing reviews on e-commerce sites.

[0279] Device: The electronic device on which a user views reviews, such as a smartphone, tablet, or computer, that is connected to the internet.

[0280] software

[0281] Natural Language Processing (NLP) technology: Software used to analyze reviews, specifically Google Cloud Natural Language and Microsoft Azure Text Analytics.

[0282] Emotion engine: Software for recognizing user emotions, including facial recognition and voice analysis technologies.

[0283] Database Management System: Software such as MongoDB or MySQL to store collected reviews and user history information.

[0284] 3. Data processing and calculation

[0285] Collecting reviews

[0286] The server collects relevant product reviews based on the search query received from the user using APIs of multiple e-commerce sites, and stores the collected reviews in an internal database.

[0287] Review analysis

[0288] The server applies NLP techniques to analyze the collected reviews. The analysis process includes the following steps:

[0289] Classifying reviews: Categorizing reviews into positive, negative, and neutral categories.

[0290] Keyphrase extraction: Extract important information and evaluation points.

[0291] Sentiment analysis: Analyze the emotional tone of each review.

[0292] filtering

[0293] The server detects and filters stealth marketing and fake reviews, performs credibility scoring, and weeds out fraudulent reviews.

[0294] Generate a summary

[0295] Generate easy-to-read summaries from reviews that have passed the credibility assessment, extracting only the most important points and summarizing them in short sentences.

[0296] Generate personalized reviews

[0297] Generates individually optimized review information based on a user's browsing and purchasing history. Analyzes past preferences and reactions to provide the most relevant information to the user.

[0298] emotion recognition

[0299] To recognize the user's emotions, the server uses an emotion engine that utilizes facial recognition and voice analysis technologies to identify the user's current emotional state.

[0300] Customized review display

[0301] Based on the user's emotional state, the server customizes the display of reviews: if the user is in a positive state, positive reviews are highlighted;

[0302] 4. Examples of concrete examples and prompts

[0303] Specific examples

[0304] When a user searches for "smartphone cases," the server collects relevant reviews from multiple e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries from reliable reviews. It then creates personalized reviews based on the user's browsing history and emotional state and displays them on the device.

[0305] Prompt Sentence Examples

[0306] If a user wants to get a detailed review summary using a generative AI model, they can enter the following prompt:

[0307] "Summarize your phone case review, especially highlighting any mentions of durability, design, and price."

[0308] In this way, the invention quickly and accurately selects and displays product reviews on a user's e-commerce site, improving the user's purchasing experience.

[0309] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0310] Step 1:

[0311] A user searches for product reviews.

[0312] Specific operation: A user accesses an e-commerce site, enters a product name or category in the search bar, for example, "smartphone case," and clicks the search button.

[0313] Input: A search query such as a product name or category.

[0314] Output: The search query information sent to the server.

[0315] Step 2:

[0316] The server collects product reviews.

[0317] Specific operation: Based on the search query received from the user, the server uses an API to collect relevant product reviews from multiple major e-commerce sites.

[0318] Input: Search query information submitted by the user.

[0319] Output: Associated product reviews stored in an internal database.

[0320] Step 3:

[0321] The server analyzes the collected reviews.

[0322] How it works: The server applies natural language processing (NLP) techniques to analyze the collected reviews, using Google Cloud Natural Language and Microsoft Azure Text Analytics. The analysis process includes review classification, key phrase extraction, and sentiment analysis.

[0323] Input: Collected product reviews.

[0324] Output: Classification results, extracted key phrases, and sentiment analysis results.

[0325] Step 4:

[0326] The server performs the trust rating and filtering.

[0327] What it does: Scores the trustworthiness of analyzed reviews and filters out stealth marketing and fake reviews. This includes the process of detecting unnatural patterns and assigning a trustworthiness score.

[0328] Input: Analysis results and review content.

[0329] Output: Trustworthy scored reviews, trustworthy reviews with fraudulent reviews filtered out.

[0330] Step 5:

[0331] The server generates an easy-to-read summary.

[0332] What it does: It generates easy-to-read summaries from reliable reviews, extracting specific evaluation points and summarizing them in short sentences. For example, it might summarize something like, "This smartphone case is durable and has a good design, but is a little pricey."

[0333] Enter: reliable reviews.

[0334] Output: A user-readable summary.

[0335] Step 6:

[0336] The server generates the personalized review.

[0337] What it does: Generates personalized reviews based on users' browsing and purchasing history, analyzing past preferences and reactions to provide users with the most relevant information.

[0338] Input: User browsing and purchasing history, trusted reviews.

[0339] Output: A personalized product review.

[0340] Step 7:

[0341] The server recognizes the user's emotions.

[0342] What it does: It uses an emotion engine to recognize the user's emotions. It uses facial recognition and voice analysis technologies to identify the user's current emotional state.

[0343] Input: User's facial image and voice data.

[0344] Output: The user's emotional state (positive, negative, etc.).

[0345] Step 8:

[0346] The server customizes the review display based on the emotional state.

[0347] Specific behavior: Based on the information obtained from the emotion engine, the summary is customized to provide review information that matches the user's emotional state. For example, if the user is in a positive mood, positive reviews will be highlighted.

[0348] Input: Emotional state information from the emotion engine, generated summary.

[0349] Output: Customized review display information.

[0350] Step 9:

[0351] Display the generated summary on your terminal.

[0352] Specific behavior: The generated summary and personalized review are displayed on the user's device. For example, a message like "Many users rate this case highly for its durability and design, and although it is a little pricey, they are very satisfied with it" is displayed on the smartphone screen.

[0353] Input: Customized review display information.

[0354] Output: Summary and review information displayed on the user's device.

[0355] (Application example 2)

[0356] 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."

[0357] Modern e-commerce sites make it difficult for users to quickly and accurately evaluate product reviews. They also contain a large number of stealth marketing and fake reviews, creating the risk of making purchasing decisions based on unreliable information. Furthermore, reviews are not personalized based on each user's preferences and emotional state, potentially reducing user satisfaction.

[0358] 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 collecting reviews based on a search query received from a user, means for analyzing the collected reviews using natural language processing technology, categorizing the reviews, and extracting key phrases, means for evaluating the credibility of the reviews and filtering out stealth marketing and false reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized summaries based on the user's browsing history and purchase history, means for analyzing the user's facial expressions using facial recognition technology and identifying their emotional state, means for analyzing the user's vocalizations using speech analysis technology and identifying their emotional state, means for customizing review displays based on information obtained from the emotion engine, and means for displaying the generated summaries on the user's terminal. This allows users to quickly and reliably obtain review information and enjoy review displays optimized based on their own emotional state and past purchase history.

[0359] A "search query" is a keyword or phrase that a user enters to search for a specific product or information.

[0360] A "review" is a piece of text that contains a user's evaluation or opinion of a product or service.

[0361] "Natural language processing technology" is a technology that allows computers to analyze, understand, and generate human language.

[0362] A "key phrase" is a word or phrase that is considered to be of high importance in a sentence.

[0363] "Stealth marketing" is a marketing technique that intentionally promotes products and services without consumers realizing it.

[0364] A "false review" is a review that contains information that is not true.

[0365] A "summary" is a short summary of the main content of the original text.

[0366] A "personalized summary" is a summary that is optimized based on an individual user's interests and past behavior.

[0367] "Facial recognition technology" is a technology that identifies and analyzes people's faces from images and videos.

[0368] "Voice analysis technology" is a technology that analyzes voice data and analyzes content, emotions, etc.

[0369] An "emotion engine" is a technology that analyzes the user's emotions and optimizes the system's operation based on the results.

[0370] This invention provides a system that enables users to quickly and easily select product reviews on an e-commerce site, and also has the function of recognizing the user's emotions and customizing the review display. Specific embodiments of this system are described below.

[0371] Generating a Program

[0372] This system is composed of a program that combines multiple methods. First, the server collects product reviews based on search queries received from users. To do this, the server retrieves relevant reviews from major e-commerce sites via API and stores them in an internal database.

[0373] Processing Overview

[0374] The following hardware and software are used to implement this program:

[0375] Hardware: Smartphone

[0376] Software: NLP analysis tools (e.g., Google Cloud Natural Language API), emotion analysis tools (e.g., Microsoft Azure Emotion API)

[0377] Database: NoSQL database (e.g. MongoDB)

[0378] Server: AWS Lambda (serverless computing)

[0379] 1. Collect reviews

[0380] The server collects reviews based on the search query submitted by the user. For example, when a user searches for "smartphone case," the server collects related reviews via the e-commerce site's API and stores them in a NoSQL database.

[0381] 2. NLP analysis

[0382] The server uses the Google Cloud Natural Language API to analyze the collected reviews, including classifying reviews as positive, negative, or neutral, extracting key phrases, and analyzing emotional tone, allowing it to quickly provide users with the information they need most.

[0383] 3. Trustworthiness Assessment and Filtering

[0384] The server evaluates the authenticity of the reviews and filters out stealth marketing and fake reviews. This step removes unusual patterns (such as a large number of posts from the same IP address) and reviews with low scores.

[0385] 4. Summary Generation

[0386] The server generates summaries from reliable reviews, which are short summaries of key features of the reviews to help users make a quick decision. For example, a summary might read, "This smartphone case is durable and has a good design, but is a bit pricey."

[0387] 5. Emotion recognition

[0388] The server uses the Microsoft Azure Emotion API to recognize the user's current emotional state. To do this, it uses facial expression analysis using the user's smartphone camera and emotion analysis technology from voice. This allows it to identify the user's emotional state.

[0389] 6. Personalized Display

[0390] The server customizes the display of reviews based on information obtained from the emotion engine. For example, when a user is in a positive emotional state, it highlights positive reviews. This function provides information optimized for each user.

[0391] 7. Display

[0392] Finally, the generated summary is displayed on the user's device. If the device is a smartphone, the appropriate summary is visually presented on the screen.

[0393] Specific examples

[0394] When a user searches for "smartphone case," the server sends the following prompt to the generative AI model:

[0395] Get product-related reviews and categorize them as positive, negative, or neutral based on sentiment. Extract key phrases and generate summaries. Eliminate stealth marketing patterns and assess trustworthiness. Display customized summaries based on the user's emotional state.

[0396] This allows users to quickly obtain reliable review information and enjoy optimized review displays based on their own emotional state and past purchasing history.

[0397] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0398] Step 1:

[0399] A user enters keywords related to a specific product name or category into the search bar and clicks the search button. This input generates a search query to retrieve reviews related to the product. The server receives this search query and secures the data to proceed to the next step.

[0400] Step 2:

[0401] The server collects relevant product reviews from major e-commerce sites based on the received search query. Specifically, it retrieves the relevant reviews via API and stores them in a NoSQL database. For example, it collects hundreds of reviews about "smartphone cases." This allows review information based on the product the user is looking for to be stored in the database.

[0402] Step 3:

[0403] The server applies natural language processing (NLP) technology to analyze the collected product reviews. First, it classifies the review content as positive, negative, or neutral. Next, it extracts important information and evaluation points (key phrases) from each review. It also performs sentiment analysis of the reviews to determine the emotional tone of the review. For example, it distinguishes between reviews that are "very satisfied" and reviews that are "slightly dissatisfied." This allows the information users need to be provided quickly.

[0404] Step 4:

[0405] The server evaluates the credibility of reviews based on the analysis results and filters out stealth marketing and fake reviews. First, it detects unnatural patterns and removes reviews posted in large numbers from the same IP address or reviews with abnormally high ratings. Next, it assigns a credibility score to each review and eliminates reviews with low scores. This step allows reliable reviews to be selected.

[0406] Step 5:

[0407] The server generates easy-to-read summaries from reviews that pass the reliability assessment. Specifically, it summarizes important evaluation points in a concise manner, allowing users to quickly obtain key information without having to read lengthy reviews. For example, a summary might be generated such as, "This smartphone case is highly durable and has a well-received design, but is a little pricey."

[0408] Step 6:

[0409] The server generates personalized summaries based on the user's browsing and purchasing history. It analyzes the types of products the user has previously purchased and provides summaries tailored to those preferences. For example, a user who has previously preferred durable products might receive a summary such as "This case is particularly durable."

[0410] Step 7:

[0411] The server uses an emotion engine to recognize the user's emotions. To do this, it analyzes the user's facial expressions using the device (smartphone) camera and identifies the user's emotional state using the Microsoft Azure Emotion API. It also analyzes the user's vocalizations to recognize the user's emotional state.

[0412] Step 8:

[0413] The server customizes the review display based on the information obtained from the emotion engine and the personalized summary. For example, if the user is in a positive state, it highlights positive reviews, and if the user is in an anxious state, it highlights reliable reviews. This step provides the user with the most appropriate information according to their emotional state.

[0414] Step 9:

[0415] The generated summary is displayed on the device (smartphone) screen. Users can quickly obtain the appropriate summary information to quickly evaluate the product. For example, a summary such as "Many users rate this case highly for its durability and design, and although the price is somewhat high, satisfaction is high" may be displayed.

[0416] 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.

[0417] 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.

[0418] 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.

[0419] [Second embodiment]

[0420] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0421] 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.

[0422] 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).

[0423] 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.

[0424] 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.

[0425] 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).

[0426] 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.

[0427] 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.

[0428] 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.

[0429] 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.

[0430] 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.

[0431] 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."

[0432] The present invention provides a system that allows users to quickly and easily select product reviews on an e-commerce site. This system is implemented using the following programs in particular.

[0433] Program processing explanation

[0434] 1. User searches for product reviews

[0435] A user visits an e-commerce site and enters keywords related to a specific product name or category into the search bar. Based on this search query, the server begins searching for relevant reviews. For example, a user enters "smartphone case" and clicks the search button.

[0436] 2. The server collects reviews

[0437] The server collects relevant product reviews from major e-commerce sites based on the search query entered by the user. The reviews are retrieved via API and stored in an internal database. This results in the collection of hundreds of reviews related to "smartphone cases."

[0438] 3. Review Analysis Using Natural Language Processing (NLP)

[0439] The server analyzes the collected reviews using natural language processing technology, which performs the following processes:

[0440] Review categorization: Categorize reviews as positive, negative, or neutral. For example, based on the content of the review, the part that says "high durability" can be categorized as positive, and conversely, the part that says "high price" can be categorized as negative.

[0441] Key phrase extraction: Extracting important information from each review, such as "nice design" or "durable and durable."

[0442] Sentiment analysis: Analyzing the tone of a review to determine emotional intensity, for example, distinguishing between "very satisfied" and "somewhat dissatisfied" reviews.

[0443] 4. Filtering and Eliminating Stealth Marketing

[0444] The server then uses the analysis results to assess trustworthiness and filter out fraudulent reviews, including:

[0445] Unnatural pattern detection: Automatically detects reviews posted in large numbers from the same IP address or reviews with abnormally high ratings.

[0446] Trustworthiness scoring: Each review is assigned a certain score to filter out fraudulent reviews. For example, reviews submitted in large numbers in a short period of time are given a low score and are subject to elimination.

[0447] 5. Generating a Summary

[0448] The server generates an easy-to-read summary based on the filtered reviews. This allows users to quickly obtain only the main points of evaluation without having to read lengthy reviews. For example, it generates a summary such as, "This smartphone case is highly durable and has a well-received design, but is a little pricey."

[0449] 6. Providing personalized reviews

[0450] The server generates personalized reviews based on the user's browsing and purchasing history. For example, if a user has previously purchased products that emphasize durability, the server provides a summary that highlights information such as, "This case is particularly durable."

[0451] 7. Display summary on terminal

[0452] The generated summary is displayed on the user's device. Summary information about the product is displayed on the screen of a smartphone or PC, allowing the user to quickly check important information. For example, a message may appear on a smartphone screen saying, "Many users rate this case highly for its durability and design, and although the price is somewhat high, satisfaction is high."

[0453] Specific examples

[0454] When a user searches for "smartphone case" on their smartphone, the server collects relevant reviews from e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries. These summaries are then displayed as personalized summaries on the user's device, allowing the user to quickly and efficiently evaluate the product.

[0455] A system based on this claim would speed up understanding of product reviews, increase their reliability, and improve the consumer buying experience.

[0456] The processing flow will be explained below.

[0457] Step 1:

[0458] A user types keywords related to a specific product name or category into the search bar of an e-commerce site, and is ready to find reviews that match their interests.

[0459] Step 2:

[0460] The server collects relevant product reviews from major e-commerce sites based on the search query received from the user. The server uses an API to retrieve reviews that match the keywords and stores them in an internal database.

[0461] Step 3:

[0462] The server starts analyzing the collected product reviews by applying natural language processing (NLP) techniques. The server performs the following processes:

[0463] Classification of reviews: Positive, negative, or neutral.

[0464] Key phrase extraction: Extract important information and evaluation points.

[0465] Sentiment analysis: Analyze the emotional tone of each review.

[0466] Step 4:

[0467] The server uses the analysis results to assess the trustworthiness of reviews and filter out stealth marketing and fake reviews. This includes:

[0468] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with only unusually high ratings.

[0469] Trustworthiness Scoring: Each review is assigned a trustworthiness score to help weed out fraudulent reviews.

[0470] Step 5:

[0471] The server generates easy-to-read summaries from reviews that have passed the credibility assessment. The server concisely extracts key opinions and evaluation points and summarizes them in a format that users can quickly understand.

[0472] Step 6:

[0473] The server generates personalized reviews based on the user's browsing and purchasing history, analyzing the user's past preferences and reactions and providing individually optimized review information.

[0474] Step 7:

[0475] The terminal displays the generated summary information to the user, who can then visually view the generated summary on the screen of their smartphone or computer, allowing them to quickly check the product evaluation.

[0476] This is the specific processing flow of the "AI Review Summarizer" system, which enables users to quickly organize and accurately understand vast amounts of review information.

[0477] Example 1

[0478] 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."

[0479] When purchasing a product on an online e-commerce site, users need to find useful information from numerous reviews, but this task requires time and effort. Furthermore, some reviews may be fraudulent or unreliable, which can lead to misjudgments. Furthermore, because different users have different priorities, it is not always easy to find the most suitable review for each user. Therefore, there is a need for a fast and reliable review selection system that can provide personalized reviews tailored to users' preferences.

[0480] 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.

[0481] In this invention, the server includes means for collecting product reviews based on a search query received from a user, means for analyzing the collected product reviews using natural language processing technology, categorizing the reviews, and extracting key phrases, means for evaluating the reliability of the reviews and filtering out fraudulent reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized summaries based on the user's browsing history and purchase history, and means for providing the generated summaries to a display device. This allows users to quickly and efficiently obtain reliable product reviews and make purchasing decisions based on information that suits their needs.

[0482] A "search query" is a keyword or phrase that a user enters to search for information related to a particular product or category.

[0483] A "product review" is a written description of the product's usage and evaluation by a user who has purchased the product on an e-commerce site or the like.

[0484] "Natural language processing technology" is a technology that enables computers to understand and analyze human language and extract its meaning.

[0485] "Review categorization" is the process of dividing collected product reviews into categories based on their evaluation, such as positive, negative, or neutral.

[0486] A "key phrase" is a major phrase in a piece of text that expresses important information or a particular theme.

[0487] "Review credibility assessment" is the process of determining how accurate and trustworthy the collected reviews are.

[0488] "Fraudulent reviews" are false or exaggerated reviews posted with the intent to manipulate product sales.

[0489] "Filtering" is the process of removing unnecessary or fraudulent information from collected data.

[0490] A "summary" is a concise summary of the main points extracted from a lengthy review.

[0491] A "personalized summary" is a summary that is individually customized based on a particular user's preferences and purchasing history.

[0492] A "display device" is a device for visually presenting the generated summary or other information to a user, including a smartphone or personal computer.

[0493] This invention provides an information processing system that allows users to quickly and easily select product reviews on e-commerce sites. The system collects relevant reviews based on a user's search query, analyzes them using natural language processing (NLP), evaluates their credibility, generates summaries, and provides personalized summaries to users.

[0494] System Configuration

[0495] The system consists of the following main components:

[0496] 1. User device: The device where a user enters a search query and sees the results. This can be a smartphone or a computer.

[0497] 2. Server: A central processing unit that receives search queries from users and collects, analyzes, filters, and summarizes reviews. The server is made up of hardware with a high-performance processor and large memory capacity.

[0498] 3. Natural Language Processing Engine: Software for analyzing reviews. NLP tools such as Python's NLTK (Natural Language Toolkit) and Gensim are used.

[0499] Data processing and calculation

[0500] The server does the following:

[0501] 1. Review collection: The server collects relevant product reviews using the APIs of major e-commerce sites based on the search query received from the user. For example, if a user searches for "smartphone case," the server retrieves data via the APIs of Amazon and Rakuten Ichiba.

[0502] 2. Review Analysis: We analyze the collected reviews using natural language processing (NLP). We classify reviews as "positive," "negative," or "neutral," extract key phrases, and perform sentiment analysis. We use the NLTK and Gensim libraries to perform this analysis.

[0503] 3. Review credibility evaluation and filtering: The server evaluates the credibility of reviews based on the analysis results and filters out fraudulent reviews. The algorithms used include unnatural pattern detection and credibility scoring.

[0504] 4. Summary Generation: The server generates summaries from the filtered reviews. It also creates personalized summaries that emphasize different points for each user. It uses the Gensim library to generate the summary text.

[0505] Specific examples

[0506] For example:

[0507] When a user searches for "smartphone case" on their smartphone, the following prompt is entered into the system:

[0508] "I want to generate summaries from reviews about smartphone cases, classify them as positive or negative, and extract key phrases."

[0509] The server receives the search query and collects review information via APIs from Amazon and Rakuten. It then analyzes the reviews using NLTK and Gensim and classifies them as positive, negative, or neutral. It then scores the reviews' trustworthiness, filters out unnatural patterns, and generates a personalized summary tailored to the user's preferences. This summary is then displayed on the browser screen of a smartphone or PC, allowing users to easily check important information about the review.

[0510] This invention enables users to quickly and efficiently obtain highly reliable product reviews and make purchasing decisions based on information that meets their needs.

[0511] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0512] Step 1:

[0513] User searches for product reviews

[0514] A user accesses an e-commerce site using a smartphone or PC and enters keywords related to a specific product name or category into the search bar. For example, if the user enters "smartphone case," the search query is sent to the server when the user clicks the search button.

[0515] Input: The search query entered by the user (e.g., "smartphone case")

[0516] Output: The search query sent to the server

[0517] Step 2:

[0518] Server collects reviews

[0519] The server uses the APIs of major e-commerce sites to collect relevant product reviews based on the search query received from the user. Secure data acquisition is performed using an API key, and the acquired review information is stored in an internal database. For example, reviews on "smartphone cases" are collected using the APIs of Amazon and Rakuten Ichiba.

[0520] Input: The search query received by the server

[0521] Output: A database of collected review information

[0522] Step 3:

[0523] Review analysis using natural language processing (NLP)

[0524] The server analyzes the collected reviews using natural language processing (NLP) technology. Specifically, it performs the following processes:

[0525] Review Classification: Classify review text as positive, negative, or neutral. For example, "The case is very sturdy" is classified as positive, and "The price is too high" is classified as negative.

[0526] Keyphrase extraction: Important keyphrases are extracted using TF-IDF and Word2Vec algorithms. Examples include "good design" and "durable."

[0527] Sentiment analysis: Analyzes the sentiment of the review and evaluates the strength of the sentiment. For example, a "Very Satisfied" review will have a high positive score, while a "Slightly Dissatisfied" review will have a negative score.

[0528] Input: Collected review information

[0529] Output: Analyzed review classification, key phrases, and sentiment analysis data

[0530] Step 4:

[0531] Review credibility rating and filtering

[0532] The server evaluates the reliability of reviews based on the analysis results and performs filtering. Specifically, it performs the following processes:

[0533] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with only unusually high ratings.

[0534] Trustworthiness Scoring: Assign a score to each review and automatically filter out fraudulent reviews with low scores.

[0535] Input: Parsed review classification, key phrases, and sentiment analysis data

[0536] Output: Filtered data containing only reliable reviews

[0537] Step 5:

[0538] Generate a summary

[0539] The server generates summaries based on the filtered reviews. Using the Gensim library, it extracts key points from the review text and summarizes them in a concise format to create easy-to-read summaries. For example, it generates a summary such as, "This smartphone case is durable and has a good design, but is a little pricey."

[0540] Input: Filtered reviews

[0541] Output: Generated summary

[0542] Step 6:

[0543] Generate personalized reviews

[0544] The server generates a personalized summary based on the user's browsing and purchasing history. For a user who has previously purchased products that emphasize durability, the server generates a summary that emphasizes the fact that "this case is particularly durable."

[0545] Input: User browsing history, purchase history, filtered reviews

[0546] Output: Personalized Summary

[0547] Step 7:

[0548] Display summary to terminal

[0549] The generated summary is displayed on the user's device. Summary information about the product is displayed on the screen of a smartphone or computer, allowing the user to quickly check important information. Specifically, the smartphone browser or dedicated app screen displays the message, "Many users rate this case highly for its durability and design, and although the price is a little high, satisfaction is high."

[0550] Input: Generated summary, personalized summary

[0551] Output: Summary information displayed on the user's terminal

[0552] (Application example 1)

[0553] 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."

[0554] Online e-commerce sites contain a huge number of product reviews, making it difficult for users to quickly access the information they need. Conventional systems lack the ability to evaluate and filter the reliability of reviews, and many contain stealth marketing and fraudulent reviews. Furthermore, personalized reviews tailored to each user's needs are not provided, making it difficult for users to effectively utilize reviews. This can lead to consumers making incorrect decisions or wasting unnecessary time when choosing products.

[0555] 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.

[0556] In this invention, the server includes means for collecting product reviews based on a search query received from a user, means for analyzing the collected product reviews using natural language processing technology to categorize the reviews and extract key phrases, means for evaluating the reliability of the reviews and filtering out stealth marketing and false reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized reviews based on the user's browsing history and purchasing history, means for displaying the generated summaries on the user's device, means for analyzing the content of the reviews and categorizing them as positive, negative, or neutral, and means for analyzing detailed information about the products viewed by the user and generating prompt sentences. This allows users to quickly access reliable review summaries and efficiently obtain information tailored to their purchasing behavior.

[0557] A "search query" is a keyword or phrase that a user enters when searching for a product or information.

[0558] A "product review" is a record of the evaluation and impressions of a product made by a user who has purchased the product.

[0559] "Natural language processing technology" is a technology that allows computers to understand, analyze, and generate human language.

[0560] "Keyphrases" refer to important words or phrases within a review or document.

[0561] "Stealth marketing" refers to advertising activities that are carried out without consumers realizing that they are advertising.

[0562] A "false review" is a review that contains false information that differs from the actual performance or usability of the product.

[0563] A "summary" is a short summary that extracts the main points from a long document or review.

[0564] A "personalized review" is a review that is customized based on the interests and preferences of an individual user.

[0565] "Evaluating credibility" refers to the act of judging and evaluating the accuracy and reliability of collected reviews.

[0566] This invention provides a system that allows users to quickly and easily select product reviews on an e-commerce site. This system is realized through the following steps.

[0567] First, a user accesses an e-commerce site from a smartphone or computer and enters a search query related to a specific product name or category. This search query is sent to a server, which then collects relevant product reviews from the e-commerce site. The collected reviews are then stored in the server's database.

[0568] The server then analyzes the collected reviews using natural language processing (NLP) technology. This analysis includes categorizing reviews into positive, negative, and neutral categories and extracting key phrases. Specific analysis techniques used include tools such as the Python nltk library and scikit-learn.

[0569] The server then evaluates the credibility of reviews based on the analysis results and filters out fraudulent reviews. This evaluation is based on review posting frequency and IP address pattern analysis. It also applies algorithms to automatically detect stealth marketing and fake reviews. This ensures users have access to reliable review information.

[0570] The server generates summaries of the analyzed and filtered reviews. These summaries are short summaries that extract the main points of evaluation from the long reviews. For example, a summary might be generated that reads, "This smartphone case is durable and has a good design, but is a little pricey."

[0571] The server also generates personalized reviews based on the user's browsing and purchasing history. These personalized reviews provide information customized to the user's preferences and interests. For example, for a user who has previously purchased a durable product, the server may emphasize that "this case is particularly durable."

[0572] Finally, the generated summary and personalized review are displayed on the user's device. Users can view the concise review information on their smartphone or computer screen, allowing them to quickly and efficiently understand product ratings and obtain information that will help them make a purchase.

[0573] For example, when a user searches for "smartphone case" on their smartphone, the server collects relevant reviews from e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates a summary, which is then displayed as a personalized summary on the user's device.

[0574] An example prompt is, "Based on the product name searched by the user, collect relevant reviews, analyze the reviews using natural language processing, and generate summaries. In particular, display them separately from the perspectives of positive, negative, and neutral."

[0575] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0576] Step 1:

[0577] A user accesses an e-commerce site using a smartphone or computer and enters a search query related to a specific product name or category. The search query is sent from the device to the server. The input data at this time is the keywords or phrases entered by the user.

[0578] Step 2:

[0579] The server collects relevant product reviews from major e-commerce sites based on the received search query. The collection is done using an API, and the acquired review information is stored in the server's database. The output data at this time is raw review information.

[0580] Step 3:

[0581] The server uses natural language processing (NLP) technology to analyze the collected product reviews. Specifically, it categorizes the reviews into positive, negative, and neutral categories and extracts important key phrases. The NLP technology used includes Python's nltk library and scikit-learn. The input data is raw review information, and the output data is analyzed review information.

[0582] Step 4:

[0583] The server evaluates the reliability of the reviews based on the analysis results and filters out fraudulent reviews. The filtering method uses an algorithm that automatically detects large numbers of posts from the same IP address and reviews with abnormally high ratings. The input data is the analyzed review information, and the output data is the reliability-evaluated review information.

[0584] Step 5:

[0585] The server generates easy-to-read summaries from the reviews that have been evaluated for trustworthiness. The summary generation uses technology to extract the main evaluation points that have been analyzed and summarize them in a short form. For example, a summary such as "It is highly durable and has a popular design, but it is expensive" may be generated. The input data at this time is the review information that has been evaluated for trustworthiness, and the output data is the summary information.

[0586] Step 6:

[0587] The server generates personalized reviews based on the user's browsing history and purchase history. This utilizes information such as the user's past product preferences. For example, for a user who prioritizes durability, the server emphasizes information such as "This case is particularly durable." The input data at this time are summary information and user history information, and the output data is personalized summary information.

[0588] Step 7:

[0589] The generated summary and personalized review are displayed on the user's device. The user can view the concisely summarized review information on the screen of a smartphone or computer. The input data is the personalized summary information, and the output data is the summary information displayed on the user's device.

[0590] Step 8:

[0591] The server analyzes detailed information about the product viewed by the user and generates a prompt sentence. This prompt sentence is used to complement the user's next search query and review confirmation. For example, a prompt sentence may be generated that reads, "Based on the product name searched by the user, please collect related reviews, analyze the reviews using natural language processing, and generate summaries. Please display them in particular from the perspectives of positive, negative, and neutral." The input data at this time are browsing history information and review information, and the output data is the prompt sentence.

[0592] 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.

[0593] The present invention is a system that enables users to quickly and easily select product reviews on e-commerce sites, and also has the ability to recognize users' emotions and customize the review display. The specific program processing will be explained in detail below, with examples.

[0594] Program processing explanation

[0595] 1. User searches for product reviews

[0596] A user visits an e-commerce site and enters keywords related to a specific product name or category into the search bar, which prepares the site to retrieve reviews tailored to the user's interests. For example, a user enters "smartphone cases" and clicks the search button.

[0597] 2. The server collects reviews

[0598] The server collects relevant product reviews from major e-commerce sites based on the search query received from the user. The reviews are retrieved via API and stored in an internal database. For example, hundreds of reviews on "smartphone cases" are collected.

[0599] 3. Review Analysis Using Natural Language Processing (NLP)

[0600] The server analyzes the collected product reviews using natural language processing (NLP) techniques. It performs the following processes:

[0601] Review categorization: Categorize reviews as positive, negative, or neutral. For example, categorize reviews that say "high durability" as positive and reviews that say "high price" as negative.

[0602] Key phrase extraction: Extract important information and evaluation points, such as "good design" or "durable and durable."

[0603] Sentiment analysis: Analyzing the emotional tone of each review, for example, distinguishing between "Very Satisfied" and "Somewhat Dissatisfied."

[0604] 4. Filtering and Eliminating Stealth Marketing

[0605] The server evaluates the credibility of the reviews based on the analysis results and filters out stealth marketing and fake reviews. This includes the following actions:

[0606] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with an unusually high number of positive ratings.

[0607] Credibility scoring: Each review is assigned a credibility score to filter out fraudulent reviews. For example, reviews posted in large numbers in a short period of time will be assigned a low score and will be subject to elimination.

[0608] 5. Generating a Summary

[0609] The server generates easy-to-read summaries from reviews that pass the trust assessment. Users can quickly obtain only the main points of evaluation without having to read lengthy reviews. For example, a summary such as "This smartphone case is highly durable and has a well-received design, but is a little pricey" can be generated.

[0610] 6. Providing personalized reviews

[0611] The server generates personalized reviews based on the user's browsing and purchasing history. It analyzes the user's past preferences and reactions and provides individually optimized review information. In particular, for users who have previously preferred products that emphasize durability, it provides a summary such as "This case is particularly durable."

[0612] 7. Emotion Recognition with Emotion Engine

[0613] The server uses an emotion engine to recognize the user's emotion. It performs the following operations:

[0614] Facial recognition technology: Analyzes the user's facial expressions to identify their current emotional state. For example, it can detect when a user is unhappy.

[0615] Voice analysis: Analyzing the user's vocal utterances to identify their emotional state, for example, by analyzing the user's tone of voice and word choice.

[0616] 8. Customize review display based on sentiment

[0617] The server customizes the summary based on the information obtained from the emotion engine, providing review information that matches the user's emotional state. For example, if the user is in a positive mood, the server highlights the same positive reviews.

[0618] 9. Display summary on terminal

[0619] The generated summary is displayed on the user's device. The device visually presents the generated summary on the screen of a smartphone or PC. For example, a smartphone screen might say, "Many users rate this case highly for its durability and design, and although the price is a little high, satisfaction is high."

[0620] Specific examples

[0621] When a user searches for "smartphone case," the server collects related reviews and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries from reliable reviews. It then creates personalized reviews based on the user's browsing history and emotional state and displays them on the device. This allows users to quickly and accurately evaluate products.

[0622] The system according to this claim enhances the consumer buying experience by speeding up the understanding of product reviews and customizing the display according to the user's sentiment.

[0623] The processing flow will be explained below.

[0624] Step 1:

[0625] A user visits an e-commerce site and types keywords related to a specific product name or category into the search bar, generating a search query to find reviews that match the user's interests.

[0626] Step 2:

[0627] The server collects related product reviews from major e-commerce sites based on the search query received from the user. The server uses an API to retrieve reviews that match the keywords and stores them in an internal database. For example, it collects reviews related to "smartphone cases."

[0628] Step 3:

[0629] The server applies natural language processing (NLP) techniques to the collected product reviews to analyze them, including:

[0630] Review classification: Classify collected reviews as positive, negative, or neutral. For example, "The case is durable" is classified as positive, and "The price is high" is classified as negative.

[0631] Key phrase extraction: Extract key information and evaluation points from each review, such as "good design" or "durable and durable."

[0632] Sentiment analysis: Analyzing the emotional tone of a review to determine its strength, for example, distinguishing between "Very Satisfied" and "Slightly Dissatisfied."

[0633] Step 4:

[0634] The server evaluates the credibility of the reviews based on the analysis results and filters out stealth marketing and fake reviews. It does the following:

[0635] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with abnormally high ratings.

[0636] Trustworthiness Scoring: Each review is assigned a trustworthiness score to filter out fraudulent reviews. For example, reviews posted in large numbers in a short period of time are filtered out as low scores.

[0637] Step 5:

[0638] The server generates easy-to-read summaries from reviews that have passed the reliability assessment. The server succinctly summarizes key opinions and evaluation points so that users can understand them quickly. For example, it generates a summary such as, "This smartphone case is highly durable and has a well-received design, but the price is a little high."

[0639] Step 6:

[0640] The server generates personalized reviews based on the user's browsing and purchasing history. This analyzes the user's past preferences and reactions and provides optimized review information. For example, a user who prioritizes durability might be provided with a summary such as "This case is particularly durable."

[0641] Step 7:

[0642] The server uses an emotion engine to recognize the user's emotions, which involves the following steps:

[0643] Facial recognition technology: Analyzes the user's facial expressions to identify their current emotional state, for example, detecting dissatisfied facial expressions.

[0644] Voice analysis: Analyzing a user's voice to identify their emotional state, for example, determining their emotions based on their tone of voice and the way they speak.

[0645] Step 8:

[0646] The server customizes summaries based on the emotional state obtained from the emotion engine, providing users with review information that matches their emotional state. For example, if the user is in a positive mood, positive reviews will be highlighted.

[0647] Step 9:

[0648] The device displays the generated summary to the user. It is presented visually on the screen of a smartphone or computer. For example, it may display information such as, "Many users rate this case highly for its durability and design, and although it is somewhat expensive, satisfaction is high."

[0649] This is the specific processing flow of the "AI Review Summarizer" system, which allows users to quickly understand reviews and obtain information customized to their sentiment.

[0650] Example 2

[0651] 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."

[0652] Conventional product review management systems on e-commerce sites make it difficult for users to efficiently find useful information from the large number of reviews. Furthermore, stealth marketing and false reviews can make it difficult for users to make accurate decisions. Furthermore, the lack of functionality to customize reviews based on users' browsing history and emotional state makes it difficult to improve the user's purchasing experience.

[0653] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes: means for collecting product reviews based on a search query received from a user; means for analyzing the collected product reviews using natural language processing technology, categorizing the reviews, and extracting key phrases; means for evaluating the credibility of the reviews and filtering out fraudulent reviews; means for generating easy-to-read summaries from the analyzed reviews; means for generating personalized reviews based on the user's browsing history and purchasing history; means including an emotion engine for recognizing the user's emotions; means for customizing review display based on the emotion recognition results; and means for displaying the generated summaries on the user's device. This allows users to quickly and accurately obtain useful review information and make accurate decisions without being influenced by stealth marketing or fake reviews. Furthermore, the review display is customized based on the user's individual browsing history and emotional state, improving the purchasing experience.

[0654] A "search query" is a word or phrase that a user enters to search for specific information.

[0655] "Product reviews" are evaluations and comments written by users who have purchased a product about the quality and performance of that product.

[0656] "Natural language processing technology" is a technology that allows programs to understand, interpret, and generate human language.

[0657] "Classification" is the act of grouping collected data based on specific criteria.

[0658] "Keyphrase extraction" is the process of extracting important information and evaluation points from a text.

[0659] A "trustworthiness rating" is a rating used to determine whether the content of a review is trustworthy.

[0660] "Filtering" is the process of removing unwanted or harmful information.

[0661] "Summary generation" is the act of shortening a long piece of text and extracting and summarizing only the main points.

[0662] "Browsing history" is a record of pages and product information that a user has previously viewed on a website or app.

[0663] "Purchase history" is a record of products purchased by a user in the past.

[0664] A "personalized review" is a product review that is customized based on a user's individual tastes and preferences.

[0665] An "emotion engine" is a technology or system for recognizing and analyzing a user's emotional state.

[0666] "Customization" refers to adjustments or changes made to suit the needs and preferences of individual users.

[0667] A "terminal" is an electronic device used by a user to display information.

[0668] The present invention provides a system that enables users to quickly and easily select product reviews on an e-commerce site, and also has a function for recognizing user emotions and customizing the review display. A specific embodiment of this system will be described below.

[0669] 1. System Overview

[0670] The system performs a series of processes that allow users to search for product reviews and display them in a customized format. The system is designed on the premise that it is operated by a server, a user's terminal, and the user.

[0671] 2. Hardware and Software Configuration

[0672] Hardware

[0673] Server: A group of computers with high-performance processing power and large storage capacity. They are responsible for collecting and analyzing reviews on e-commerce sites.

[0674] Device: The electronic device on which a user views reviews, such as a smartphone, tablet, or computer, that is connected to the internet.

[0675] software

[0676] Natural Language Processing (NLP) technology: Software used to analyze reviews, specifically Google Cloud Natural Language and Microsoft Azure Text Analytics.

[0677] Emotion engine: Software for recognizing user emotions, including facial recognition and voice analysis technologies.

[0678] Database Management System: Software such as MongoDB or MySQL to store collected reviews and user history information.

[0679] 3. Data processing and calculation

[0680] Collecting reviews

[0681] The server collects relevant product reviews based on the search query received from the user using APIs of multiple e-commerce sites, and stores the collected reviews in an internal database.

[0682] Review analysis

[0683] The server applies NLP techniques to analyze the collected reviews. The analysis process includes the following steps:

[0684] Classifying reviews: Categorizing reviews into positive, negative, and neutral categories.

[0685] Keyphrase extraction: Extract important information and evaluation points.

[0686] Sentiment analysis: Analyze the emotional tone of each review.

[0687] filtering

[0688] The server detects and filters stealth marketing and fake reviews, performs credibility scoring, and weeds out fraudulent reviews.

[0689] Generate a summary

[0690] Generate easy-to-read summaries from reviews that have passed the credibility assessment, extracting only the most important points and summarizing them in short sentences.

[0691] Generate personalized reviews

[0692] Generates individually optimized review information based on a user's browsing and purchasing history. Analyzes past preferences and reactions to provide the most relevant information to the user.

[0693] emotion recognition

[0694] To recognize the user's emotions, the server uses an emotion engine that utilizes facial recognition and voice analysis technologies to identify the user's current emotional state.

[0695] Customized review display

[0696] Based on the user's emotional state, the server customizes the display of reviews: if the user is in a positive state, positive reviews are highlighted;

[0697] 4. Examples of concrete examples and prompts

[0698] Specific examples

[0699] When a user searches for "smartphone cases," the server collects relevant reviews from multiple e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries from reliable reviews. It then creates personalized reviews based on the user's browsing history and emotional state and displays them on the device.

[0700] Prompt Sentence Examples

[0701] If a user wants to get a detailed review summary using a generative AI model, they can enter the following prompt:

[0702] "Summarize your phone case review, especially highlighting any mentions of durability, design, and price."

[0703] In this way, the invention quickly and accurately selects and displays product reviews on a user's e-commerce site, improving the user's purchasing experience.

[0704] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0705] Step 1:

[0706] A user searches for product reviews.

[0707] Specific operation: A user accesses an e-commerce site, enters a product name or category in the search bar, for example, "smartphone case," and clicks the search button.

[0708] Input: A search query such as a product name or category.

[0709] Output: The search query information sent to the server.

[0710] Step 2:

[0711] The server collects product reviews.

[0712] Specific operation: Based on the search query received from the user, the server uses an API to collect relevant product reviews from multiple major e-commerce sites.

[0713] Input: Search query information submitted by the user.

[0714] Output: Associated product reviews stored in an internal database.

[0715] Step 3:

[0716] The server analyzes the collected reviews.

[0717] How it works: The server applies natural language processing (NLP) techniques to analyze the collected reviews, using Google Cloud Natural Language and Microsoft Azure Text Analytics. The analysis process includes review classification, key phrase extraction, and sentiment analysis.

[0718] Input: Collected product reviews.

[0719] Output: Classification results, extracted key phrases, and sentiment analysis results.

[0720] Step 4:

[0721] The server performs the trust rating and filtering.

[0722] What it does: Scores the trustworthiness of analyzed reviews and filters out stealth marketing and fake reviews. This includes the process of detecting unnatural patterns and assigning a trustworthiness score.

[0723] Input: Analysis results and review content.

[0724] Output: Trustworthy scored reviews, trustworthy reviews with fraudulent reviews filtered out.

[0725] Step 5:

[0726] The server generates an easy-to-read summary.

[0727] What it does: It generates easy-to-read summaries from reliable reviews, extracting specific evaluation points and summarizing them in short sentences. For example, it might summarize something like, "This smartphone case is durable and has a good design, but is a little pricey."

[0728] Enter: reliable reviews.

[0729] Output: A user-readable summary.

[0730] Step 6:

[0731] The server generates the personalized review.

[0732] What it does: Generates personalized reviews based on users' browsing and purchasing history, analyzing past preferences and reactions to provide users with the most relevant information.

[0733] Input: User browsing and purchasing history, trusted reviews.

[0734] Output: A personalized product review.

[0735] Step 7:

[0736] The server recognizes the user's emotions.

[0737] What it does: It uses an emotion engine to recognize the user's emotions. It uses facial recognition and voice analysis technologies to identify the user's current emotional state.

[0738] Input: User's facial image and voice data.

[0739] Output: The user's emotional state (positive, negative, etc.).

[0740] Step 8:

[0741] The server customizes the review display based on the emotional state.

[0742] Specific behavior: Based on the information obtained from the emotion engine, the summary is customized to provide review information that matches the user's emotional state. For example, if the user is in a positive mood, positive reviews will be highlighted.

[0743] Input: Emotional state information from the emotion engine, generated summary.

[0744] Output: Customized review display information.

[0745] Step 9:

[0746] Display the generated summary on your terminal.

[0747] Specific behavior: The generated summary and personalized review are displayed on the user's device. For example, a message like "Many users rate this case highly for its durability and design, and although it is a little pricey, they are very satisfied with it" is displayed on the smartphone screen.

[0748] Input: Customized review display information.

[0749] Output: Summary and review information displayed on the user's device.

[0750] (Application example 2)

[0751] 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."

[0752] Modern e-commerce sites make it difficult for users to quickly and accurately evaluate product reviews. They also contain a large number of stealth marketing and fake reviews, creating the risk of making purchasing decisions based on unreliable information. Furthermore, reviews are not personalized based on each user's preferences and emotional state, potentially reducing user satisfaction.

[0753] 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 collecting reviews based on a search query received from a user, means for analyzing the collected reviews using natural language processing technology, categorizing the reviews, and extracting key phrases, means for evaluating the credibility of the reviews and filtering out stealth marketing and false reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized summaries based on the user's browsing history and purchase history, means for analyzing the user's facial expressions using facial recognition technology and identifying their emotional state, means for analyzing the user's vocalizations using speech analysis technology and identifying their emotional state, means for customizing review displays based on information obtained from the emotion engine, and means for displaying the generated summaries on the user's terminal. This allows users to quickly and reliably obtain review information and enjoy review displays optimized based on their own emotional state and past purchase history.

[0754] A "search query" is a keyword or phrase that a user enters to search for a specific product or information.

[0755] A "review" is a piece of text that contains a user's evaluation or opinion of a product or service.

[0756] "Natural language processing technology" is a technology that allows computers to analyze, understand, and generate human language.

[0757] A "key phrase" is a word or phrase that is considered to be of high importance in a sentence.

[0758] "Stealth marketing" is a marketing technique that intentionally promotes products and services without consumers realizing it.

[0759] A "false review" is a review that contains information that is not true.

[0760] A "summary" is a short summary of the main content of the original text.

[0761] A "personalized summary" is a summary that is optimized based on an individual user's interests and past behavior.

[0762] "Facial recognition technology" is a technology that identifies and analyzes people's faces from images and videos.

[0763] "Voice analysis technology" is a technology that analyzes voice data and analyzes content, emotions, etc.

[0764] An "emotion engine" is a technology that analyzes the user's emotions and optimizes the system's operation based on the results.

[0765] This invention provides a system that enables users to quickly and easily select product reviews on an e-commerce site, and also has the function of recognizing the user's emotions and customizing the review display. Specific embodiments of this system are described below.

[0766] Generating a Program

[0767] This system is composed of a program that combines multiple methods. First, the server collects product reviews based on search queries received from users. To do this, the server retrieves relevant reviews from major e-commerce sites via API and stores them in an internal database.

[0768] Processing Overview

[0769] The following hardware and software are used to implement this program:

[0770] Hardware: Smartphone

[0771] Software: NLP analysis tools (e.g., Google Cloud Natural Language API), emotion analysis tools (e.g., Microsoft Azure Emotion API)

[0772] Database: NoSQL database (e.g. MongoDB)

[0773] Server: AWS Lambda (serverless computing)

[0774] 1. Collect reviews

[0775] The server collects reviews based on the search query submitted by the user. For example, when a user searches for "smartphone case," the server collects related reviews via the e-commerce site's API and stores them in a NoSQL database.

[0776] 2. NLP analysis

[0777] The server uses the Google Cloud Natural Language API to analyze the collected reviews, including classifying reviews as positive, negative, or neutral, extracting key phrases, and analyzing emotional tone, allowing it to quickly provide users with the information they need most.

[0778] 3. Trustworthiness Assessment and Filtering

[0779] The server evaluates the authenticity of the reviews and filters out stealth marketing and fake reviews. This step removes unusual patterns (such as a large number of posts from the same IP address) and reviews with low scores.

[0780] 4. Summary Generation

[0781] The server generates summaries from reliable reviews, which are short summaries of key features of the reviews to help users make a quick decision. For example, a summary might read, "This smartphone case is durable and has a good design, but is a bit pricey."

[0782] 5. Emotion recognition

[0783] The server uses the Microsoft Azure Emotion API to recognize the user's current emotional state. To do this, it uses facial expression analysis using the user's smartphone camera and emotion analysis technology from voice. This allows it to identify the user's emotional state.

[0784] 6. Personalized Display

[0785] The server customizes the display of reviews based on information obtained from the emotion engine. For example, when a user is in a positive emotional state, it highlights positive reviews. This function provides information optimized for each user.

[0786] 7. Display

[0787] Finally, the generated summary is displayed on the user's device. If the device is a smartphone, the appropriate summary is visually presented on the screen.

[0788] Specific examples

[0789] When a user searches for "smartphone case," the server sends the following prompt to the generative AI model:

[0790] Get product-related reviews and categorize them as positive, negative, or neutral based on sentiment. Extract key phrases and generate summaries. Eliminate stealth marketing patterns and assess trustworthiness. Display customized summaries based on the user's emotional state.

[0791] This allows users to quickly obtain reliable review information and enjoy optimized review displays based on their own emotional state and past purchasing history.

[0792] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0793] Step 1:

[0794] A user enters keywords related to a specific product name or category into the search bar and clicks the search button. This input generates a search query to retrieve reviews related to the product. The server receives this search query and secures the data to proceed to the next step.

[0795] Step 2:

[0796] The server collects relevant product reviews from major e-commerce sites based on the received search query. Specifically, it retrieves the relevant reviews via API and stores them in a NoSQL database. For example, it collects hundreds of reviews about "smartphone cases." This allows review information based on the product the user is looking for to be stored in the database.

[0797] Step 3:

[0798] The server applies natural language processing (NLP) technology to analyze the collected product reviews. First, it classifies the review content as positive, negative, or neutral. Next, it extracts important information and evaluation points (key phrases) from each review. It also performs sentiment analysis of the reviews to determine the emotional tone of the review. For example, it distinguishes between reviews that are "very satisfied" and reviews that are "slightly dissatisfied." This allows the information users need to be provided quickly.

[0799] Step 4:

[0800] The server evaluates the credibility of reviews based on the analysis results and filters out stealth marketing and fake reviews. First, it detects unnatural patterns and removes reviews posted in large numbers from the same IP address or reviews with abnormally high ratings. Next, it assigns a credibility score to each review and eliminates reviews with low scores. This step allows reliable reviews to be selected.

[0801] Step 5:

[0802] The server generates easy-to-read summaries from reviews that pass the reliability assessment. Specifically, it summarizes important evaluation points in a concise manner, allowing users to quickly obtain key information without having to read lengthy reviews. For example, a summary might be generated such as, "This smartphone case is highly durable and has a well-received design, but is a little pricey."

[0803] Step 6:

[0804] The server generates personalized summaries based on the user's browsing and purchasing history. It analyzes the types of products the user has previously purchased and provides summaries tailored to those preferences. For example, a user who has previously preferred durable products might receive a summary such as "This case is particularly durable."

[0805] Step 7:

[0806] The server uses an emotion engine to recognize the user's emotions. To do this, it analyzes the user's facial expressions using the device (smartphone) camera and identifies the user's emotional state using the Microsoft Azure Emotion API. It also analyzes the user's vocalizations to recognize the user's emotional state.

[0807] Step 8:

[0808] The server customizes the review display based on the information obtained from the emotion engine and the personalized summary. For example, if the user is in a positive state, it highlights positive reviews, and if the user is in an anxious state, it highlights reliable reviews. This step provides the user with the most appropriate information according to their emotional state.

[0809] Step 9:

[0810] The generated summary is displayed on the device (smartphone) screen. Users can quickly obtain the appropriate summary information to quickly evaluate the product. For example, a summary such as "Many users rate this case highly for its durability and design, and although the price is somewhat high, satisfaction is high" may be displayed.

[0811] 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.

[0812] 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.

[0813] 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.

[0814] [Third embodiment]

[0815] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0816] 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.

[0817] 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).

[0818] 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.

[0819] 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.

[0820] 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).

[0821] 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.

[0822] 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.

[0823] 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.

[0824] 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.

[0825] 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.

[0826] 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."

[0827] The present invention provides a system that allows users to quickly and easily select product reviews on an e-commerce site. This system is implemented using the following programs in particular.

[0828] Program processing explanation

[0829] 1. User searches for product reviews

[0830] A user visits an e-commerce site and enters keywords related to a specific product name or category into the search bar. Based on this search query, the server begins searching for relevant reviews. For example, a user enters "smartphone case" and clicks the search button.

[0831] 2. The server collects reviews

[0832] The server collects relevant product reviews from major e-commerce sites based on the search query entered by the user. The reviews are retrieved via API and stored in an internal database. This results in the collection of hundreds of reviews related to "smartphone cases."

[0833] 3. Review Analysis Using Natural Language Processing (NLP)

[0834] The server analyzes the collected reviews using natural language processing technology, which performs the following processes:

[0835] Review categorization: Categorize reviews as positive, negative, or neutral. For example, based on the content of the review, the part that says "high durability" can be categorized as positive, and conversely, the part that says "high price" can be categorized as negative.

[0836] Key phrase extraction: Extracting important information from each review, such as "nice design" or "durable and durable."

[0837] Sentiment analysis: Analyzing the tone of a review to determine emotional intensity, for example, distinguishing between "very satisfied" and "somewhat dissatisfied" reviews.

[0838] 4. Filtering and Eliminating Stealth Marketing

[0839] The server then uses the analysis results to assess trustworthiness and filter out fraudulent reviews, including:

[0840] Unnatural pattern detection: Automatically detects reviews posted in large numbers from the same IP address or reviews with abnormally high ratings.

[0841] Trustworthiness scoring: Each review is assigned a certain score to filter out fraudulent reviews. For example, reviews submitted in large numbers in a short period of time are given a low score and are subject to elimination.

[0842] 5. Generating a Summary

[0843] The server generates an easy-to-read summary based on the filtered reviews. This allows users to quickly obtain only the main points of evaluation without having to read lengthy reviews. For example, it generates a summary such as, "This smartphone case is highly durable and has a well-received design, but is a little pricey."

[0844] 6. Providing personalized reviews

[0845] The server generates personalized reviews based on the user's browsing and purchasing history. For example, if a user has previously purchased products that emphasize durability, the server provides a summary that highlights information such as, "This case is particularly durable."

[0846] 7. Display summary on terminal

[0847] The generated summary is displayed on the user's device. Summary information about the product is displayed on the screen of a smartphone or PC, allowing the user to quickly check important information. For example, a message may appear on a smartphone screen saying, "Many users rate this case highly for its durability and design, and although the price is somewhat high, satisfaction is high."

[0848] Specific examples

[0849] When a user searches for "smartphone case" on their smartphone, the server collects relevant reviews from e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries. These summaries are then displayed as personalized summaries on the user's device, allowing the user to quickly and efficiently evaluate the product.

[0850] A system based on this claim would speed up understanding of product reviews, increase their reliability, and improve the consumer buying experience.

[0851] The processing flow will be explained below.

[0852] Step 1:

[0853] A user types keywords related to a specific product name or category into the search bar of an e-commerce site, and is ready to find reviews that match their interests.

[0854] Step 2:

[0855] The server collects relevant product reviews from major e-commerce sites based on the search query received from the user. The server uses an API to retrieve reviews that match the keywords and stores them in an internal database.

[0856] Step 3:

[0857] The server starts analyzing the collected product reviews by applying natural language processing (NLP) techniques. The server performs the following processes:

[0858] Classification of reviews: Positive, negative, or neutral.

[0859] Key phrase extraction: Extract important information and evaluation points.

[0860] Sentiment analysis: Analyze the emotional tone of each review.

[0861] Step 4:

[0862] The server uses the analysis results to assess the trustworthiness of reviews and filter out stealth marketing and fake reviews. This includes:

[0863] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with only unusually high ratings.

[0864] Trustworthiness Scoring: Each review is assigned a trustworthiness score to help weed out fraudulent reviews.

[0865] Step 5:

[0866] The server generates easy-to-read summaries from reviews that have passed the credibility assessment. The server concisely extracts key opinions and evaluation points and summarizes them in a format that users can quickly understand.

[0867] Step 6:

[0868] The server generates personalized reviews based on the user's browsing and purchasing history, analyzing the user's past preferences and reactions and providing individually optimized review information.

[0869] Step 7:

[0870] The terminal displays the generated summary information to the user, who can then visually view the generated summary on the screen of their smartphone or computer, allowing them to quickly check the product evaluation.

[0871] This is the specific processing flow of the "AI Review Summarizer" system, which enables users to quickly organize and accurately understand vast amounts of review information.

[0872] Example 1

[0873] 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."

[0874] When purchasing a product on an online e-commerce site, users need to find useful information from numerous reviews, but this task requires time and effort. Furthermore, some reviews may be fraudulent or unreliable, which can lead to misjudgments. Furthermore, because different users have different priorities, it is not always easy to find the most suitable review for each user. Therefore, there is a need for a fast and reliable review selection system that can provide personalized reviews tailored to users' preferences.

[0875] 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.

[0876] In this invention, the server includes means for collecting product reviews based on a search query received from a user, means for analyzing the collected product reviews using natural language processing technology, categorizing the reviews, and extracting key phrases, means for evaluating the reliability of the reviews and filtering out fraudulent reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized summaries based on the user's browsing history and purchase history, and means for providing the generated summaries to a display device. This allows users to quickly and efficiently obtain reliable product reviews and make purchasing decisions based on information that suits their needs.

[0877] A "search query" is a keyword or phrase that a user enters to search for information related to a particular product or category.

[0878] A "product review" is a written description of the product's usage and evaluation by a user who has purchased the product on an e-commerce site or the like.

[0879] "Natural language processing technology" is a technology that enables computers to understand and analyze human language and extract its meaning.

[0880] "Review categorization" is the process of dividing collected product reviews into categories based on their evaluation, such as positive, negative, or neutral.

[0881] A "key phrase" is a major phrase in a piece of text that expresses important information or a particular theme.

[0882] "Review credibility assessment" is the process of determining how accurate and trustworthy the collected reviews are.

[0883] "Fraudulent reviews" are false or exaggerated reviews posted with the intent to manipulate product sales.

[0884] "Filtering" is the process of removing unnecessary or fraudulent information from collected data.

[0885] A "summary" is a concise summary of the main points extracted from a lengthy review.

[0886] A "personalized summary" is a summary that is individually customized based on a particular user's preferences and purchasing history.

[0887] A "display device" is a device for visually presenting the generated summary or other information to a user, including a smartphone or personal computer.

[0888] This invention provides an information processing system that allows users to quickly and easily select product reviews on e-commerce sites. The system collects relevant reviews based on a user's search query, analyzes them using natural language processing (NLP), evaluates their credibility, generates summaries, and provides personalized summaries to users.

[0889] System Configuration

[0890] The system consists of the following main components:

[0891] 1. User device: The device where a user enters a search query and sees the results. This can be a smartphone or a computer.

[0892] 2. Server: A central processing unit that receives search queries from users and collects, analyzes, filters, and summarizes reviews. The server is made up of hardware with a high-performance processor and large memory capacity.

[0893] 3. Natural Language Processing Engine: Software for analyzing reviews. NLP tools such as Python's NLTK (Natural Language Toolkit) and Gensim are used.

[0894] Data processing and calculation

[0895] The server does the following:

[0896] 1. Review collection: The server collects relevant product reviews using the APIs of major e-commerce sites based on the search query received from the user. For example, if a user searches for "smartphone case," the server retrieves data via the APIs of Amazon and Rakuten Ichiba.

[0897] 2. Review Analysis: We analyze the collected reviews using natural language processing (NLP). We classify reviews as "positive," "negative," or "neutral," extract key phrases, and perform sentiment analysis. We use the NLTK and Gensim libraries to perform this analysis.

[0898] 3. Review credibility evaluation and filtering: The server evaluates the credibility of reviews based on the analysis results and filters out fraudulent reviews. The algorithms used include unnatural pattern detection and credibility scoring.

[0899] 4. Summary Generation: The server generates summaries from the filtered reviews. It also creates personalized summaries that emphasize different points for each user. It uses the Gensim library to generate the summary text.

[0900] Specific examples

[0901] For example:

[0902] When a user searches for "smartphone case" on their smartphone, the following prompt is entered into the system:

[0903] "I want to generate summaries from reviews about smartphone cases, classify them as positive or negative, and extract key phrases."

[0904] The server receives the search query and collects review information via APIs from Amazon and Rakuten. It then analyzes the reviews using NLTK and Gensim and classifies them as positive, negative, or neutral. It then scores the reviews' trustworthiness, filters out unnatural patterns, and generates a personalized summary tailored to the user's preferences. This summary is then displayed on the browser screen of a smartphone or PC, allowing users to easily check important information about the review.

[0905] This invention enables users to quickly and efficiently obtain highly reliable product reviews and make purchasing decisions based on information that meets their needs.

[0906] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0907] Step 1:

[0908] User searches for product reviews

[0909] A user accesses an e-commerce site using a smartphone or PC and enters keywords related to a specific product name or category into the search bar. For example, if the user enters "smartphone case," the search query is sent to the server when the user clicks the search button.

[0910] Input: The search query entered by the user (e.g., "smartphone case")

[0911] Output: The search query sent to the server

[0912] Step 2:

[0913] Server collects reviews

[0914] The server uses the APIs of major e-commerce sites to collect relevant product reviews based on the search query received from the user. Secure data acquisition is performed using an API key, and the acquired review information is stored in an internal database. For example, reviews on "smartphone cases" are collected using the APIs of Amazon and Rakuten Ichiba.

[0915] Input: The search query received by the server

[0916] Output: A database of collected review information

[0917] Step 3:

[0918] Review analysis using natural language processing (NLP)

[0919] The server analyzes the collected reviews using natural language processing (NLP) technology. Specifically, it performs the following processes:

[0920] Review Classification: Classify review text as positive, negative, or neutral. For example, "The case is very sturdy" is classified as positive, and "The price is too high" is classified as negative.

[0921] Keyphrase extraction: Important keyphrases are extracted using TF-IDF and Word2Vec algorithms. Examples include "good design" and "durable."

[0922] Sentiment analysis: Analyzes the sentiment of the review and evaluates the strength of the sentiment. For example, a "Very Satisfied" review will have a high positive score, while a "Slightly Dissatisfied" review will have a negative score.

[0923] Input: Collected review information

[0924] Output: Analyzed review classification, key phrases, and sentiment analysis data

[0925] Step 4:

[0926] Review credibility rating and filtering

[0927] The server evaluates the reliability of reviews based on the analysis results and performs filtering. Specifically, it performs the following processes:

[0928] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with only unusually high ratings.

[0929] Trustworthiness Scoring: Assign a score to each review and automatically filter out fraudulent reviews with low scores.

[0930] Input: Parsed review classification, key phrases, and sentiment analysis data

[0931] Output: Filtered data containing only reliable reviews

[0932] Step 5:

[0933] Generate a summary

[0934] The server generates summaries based on the filtered reviews. Using the Gensim library, it extracts key points from the review text and summarizes them in a concise format to create easy-to-read summaries. For example, it generates a summary such as, "This smartphone case is durable and has a good design, but is a little pricey."

[0935] Input: Filtered reviews

[0936] Output: Generated summary

[0937] Step 6:

[0938] Generate personalized reviews

[0939] The server generates a personalized summary based on the user's browsing and purchasing history. For a user who has previously purchased products that emphasize durability, the server generates a summary that emphasizes the fact that "this case is particularly durable."

[0940] Input: User browsing history, purchase history, filtered reviews

[0941] Output: Personalized Summary

[0942] Step 7:

[0943] Display summary to terminal

[0944] The generated summary is displayed on the user's device. Summary information about the product is displayed on the screen of a smartphone or computer, allowing the user to quickly check important information. Specifically, the smartphone browser or dedicated app screen displays the message, "Many users rate this case highly for its durability and design, and although the price is a little high, satisfaction is high."

[0945] Input: Generated summary, personalized summary

[0946] Output: Summary information displayed on the user's terminal

[0947] (Application example 1)

[0948] 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."

[0949] Online e-commerce sites contain a huge number of product reviews, making it difficult for users to quickly access the information they need. Conventional systems lack the ability to evaluate and filter the reliability of reviews, and many contain stealth marketing and fraudulent reviews. Furthermore, personalized reviews tailored to each user's needs are not provided, making it difficult for users to effectively utilize reviews. This can lead to consumers making incorrect decisions or wasting unnecessary time when choosing products.

[0950] 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.

[0951] In this invention, the server includes means for collecting product reviews based on a search query received from a user, means for analyzing the collected product reviews using natural language processing technology to categorize the reviews and extract key phrases, means for evaluating the reliability of the reviews and filtering out stealth marketing and false reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized reviews based on the user's browsing history and purchasing history, means for displaying the generated summaries on the user's device, means for analyzing the content of the reviews and categorizing them as positive, negative, or neutral, and means for analyzing detailed information about the products viewed by the user and generating prompt sentences. This allows users to quickly access reliable review summaries and efficiently obtain information tailored to their purchasing behavior.

[0952] A "search query" is a keyword or phrase that a user enters when searching for a product or information.

[0953] A "product review" is a record of the evaluation and impressions of a product made by a user who has purchased the product.

[0954] "Natural language processing technology" is a technology that allows computers to understand, analyze, and generate human language.

[0955] "Keyphrases" refer to important words or phrases within a review or document.

[0956] "Stealth marketing" refers to advertising activities that are carried out without consumers realizing that they are advertising.

[0957] A "false review" is a review that contains false information that differs from the actual performance or usability of the product.

[0958] A "summary" is a short summary that extracts the main points from a long document or review.

[0959] A "personalized review" is a review that is customized based on the interests and preferences of an individual user.

[0960] "Evaluating credibility" refers to the act of judging and evaluating the accuracy and reliability of collected reviews.

[0961] This invention provides a system that allows users to quickly and easily select product reviews on an e-commerce site. This system is realized through the following steps.

[0962] First, a user accesses an e-commerce site from a smartphone or computer and enters a search query related to a specific product name or category. This search query is sent to a server, which then collects relevant product reviews from the e-commerce site. The collected reviews are then stored in the server's database.

[0963] The server then analyzes the collected reviews using natural language processing (NLP) technology. This analysis includes categorizing reviews into positive, negative, and neutral categories and extracting key phrases. Specific analysis techniques used include tools such as the Python nltk library and scikit-learn.

[0964] The server then evaluates the credibility of reviews based on the analysis results and filters out fraudulent reviews. This evaluation is based on review posting frequency and IP address pattern analysis. It also applies algorithms to automatically detect stealth marketing and fake reviews. This ensures users have access to reliable review information.

[0965] The server generates summaries of the analyzed and filtered reviews. These summaries are short summaries that extract the main points of evaluation from the long reviews. For example, a summary might be generated that reads, "This smartphone case is durable and has a good design, but is a little pricey."

[0966] The server also generates personalized reviews based on the user's browsing and purchasing history. These personalized reviews provide information customized to the user's preferences and interests. For example, for a user who has previously purchased a durable product, the server may emphasize that "this case is particularly durable."

[0967] Finally, the generated summary and personalized review are displayed on the user's device. Users can view the concise review information on their smartphone or computer screen, allowing them to quickly and efficiently understand product ratings and obtain information that will help them make a purchase.

[0968] For example, when a user searches for "smartphone case" on their smartphone, the server collects relevant reviews from e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates a summary, which is then displayed as a personalized summary on the user's device.

[0969] An example prompt is, "Based on the product name searched by the user, collect relevant reviews, analyze the reviews using natural language processing, and generate summaries. In particular, display them separately from the perspectives of positive, negative, and neutral."

[0970] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0971] Step 1:

[0972] A user accesses an e-commerce site using a smartphone or computer and enters a search query related to a specific product name or category. The search query is sent from the device to the server. The input data at this time is the keywords or phrases entered by the user.

[0973] Step 2:

[0974] The server collects relevant product reviews from major e-commerce sites based on the received search query. The collection is done using an API, and the acquired review information is stored in the server's database. The output data at this time is raw review information.

[0975] Step 3:

[0976] The server uses natural language processing (NLP) technology to analyze the collected product reviews. Specifically, it categorizes the reviews into positive, negative, and neutral categories and extracts important key phrases. The NLP technology used includes Python's nltk library and scikit-learn. The input data is raw review information, and the output data is analyzed review information.

[0977] Step 4:

[0978] The server evaluates the reliability of the reviews based on the analysis results and filters out fraudulent reviews. The filtering method uses an algorithm that automatically detects large numbers of posts from the same IP address and reviews with abnormally high ratings. The input data is the analyzed review information, and the output data is the reliability-evaluated review information.

[0979] Step 5:

[0980] The server generates easy-to-read summaries from the reviews that have been evaluated for trustworthiness. The summary generation uses technology to extract the main evaluation points that have been analyzed and summarize them in a short form. For example, a summary such as "It is highly durable and has a popular design, but it is expensive" may be generated. The input data at this time is the review information that has been evaluated for trustworthiness, and the output data is the summary information.

[0981] Step 6:

[0982] The server generates personalized reviews based on the user's browsing history and purchase history. This utilizes information such as the user's past product preferences. For example, for a user who prioritizes durability, the server emphasizes information such as "This case is particularly durable." The input data at this time are summary information and user history information, and the output data is personalized summary information.

[0983] Step 7:

[0984] The generated summary and personalized review are displayed on the user's device. The user can view the concisely summarized review information on the screen of a smartphone or computer. The input data is the personalized summary information, and the output data is the summary information displayed on the user's device.

[0985] Step 8:

[0986] The server analyzes detailed information about the product viewed by the user and generates a prompt sentence. This prompt sentence is used to complement the user's next search query and review confirmation. For example, a prompt sentence may be generated that reads, "Based on the product name searched by the user, please collect related reviews, analyze the reviews using natural language processing, and generate summaries. Please display them in particular from the perspectives of positive, negative, and neutral." The input data at this time are browsing history information and review information, and the output data is the prompt sentence.

[0987] 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.

[0988] The present invention is a system that enables users to quickly and easily select product reviews on e-commerce sites, and also has the ability to recognize users' emotions and customize the review display. The specific program processing will be explained in detail below, with examples.

[0989] Program processing explanation

[0990] 1. User searches for product reviews

[0991] A user visits an e-commerce site and enters keywords related to a specific product name or category into the search bar, which prepares the site to retrieve reviews tailored to the user's interests. For example, a user enters "smartphone cases" and clicks the search button.

[0992] 2. The server collects reviews

[0993] The server collects relevant product reviews from major e-commerce sites based on the search query received from the user. The reviews are retrieved via API and stored in an internal database. For example, hundreds of reviews on "smartphone cases" are collected.

[0994] 3. Review Analysis Using Natural Language Processing (NLP)

[0995] The server analyzes the collected product reviews using natural language processing (NLP) techniques. It performs the following processes:

[0996] Review categorization: Categorize reviews as positive, negative, or neutral. For example, categorize reviews that say "high durability" as positive and reviews that say "high price" as negative.

[0997] Key phrase extraction: Extract important information and evaluation points, such as "good design" or "durable and durable."

[0998] Sentiment analysis: Analyzing the emotional tone of each review, for example, distinguishing between "Very Satisfied" and "Somewhat Dissatisfied."

[0999] 4. Filtering and Eliminating Stealth Marketing

[1000] The server evaluates the credibility of the reviews based on the analysis results and filters out stealth marketing and fake reviews. This includes the following actions:

[1001] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with an unusually high number of positive ratings.

[1002] Credibility scoring: Each review is assigned a credibility score to filter out fraudulent reviews. For example, reviews posted in large numbers in a short period of time will be assigned a low score and will be subject to elimination.

[1003] 5. Generating a Summary

[1004] The server generates easy-to-read summaries from reviews that pass the trust assessment. Users can quickly obtain only the main points of evaluation without having to read lengthy reviews. For example, a summary such as "This smartphone case is highly durable and has a well-received design, but is a little pricey" can be generated.

[1005] 6. Providing personalized reviews

[1006] The server generates personalized reviews based on the user's browsing and purchasing history. It analyzes the user's past preferences and reactions and provides individually optimized review information. In particular, for users who have previously preferred products that emphasize durability, it provides a summary such as "This case is particularly durable."

[1007] 7. Emotion Recognition with Emotion Engine

[1008] The server uses an emotion engine to recognize the user's emotion. It performs the following operations:

[1009] Facial recognition technology: Analyzes the user's facial expressions to identify their current emotional state. For example, it can detect when a user is unhappy.

[1010] Voice analysis: Analyzing the user's vocal utterances to identify their emotional state, for example, by analyzing the user's tone of voice and word choice.

[1011] 8. Customize review display based on sentiment

[1012] The server customizes the summary based on the information obtained from the emotion engine, providing review information that matches the user's emotional state. For example, if the user is in a positive mood, the server highlights the same positive reviews.

[1013] 9. Display summary on terminal

[1014] The generated summary is displayed on the user's device. The device visually presents the generated summary on the screen of a smartphone or PC. For example, a smartphone screen might say, "Many users rate this case highly for its durability and design, and although the price is a little high, satisfaction is high."

[1015] Specific examples

[1016] When a user searches for "smartphone case," the server collects related reviews and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries from reliable reviews. It then creates personalized reviews based on the user's browsing history and emotional state and displays them on the device. This allows users to quickly and accurately evaluate products.

[1017] The system according to this claim enhances the consumer buying experience by speeding up the understanding of product reviews and customizing the display according to the user's sentiment.

[1018] The processing flow will be explained below.

[1019] Step 1:

[1020] A user visits an e-commerce site and types keywords related to a specific product name or category into the search bar, generating a search query to find reviews that match the user's interests.

[1021] Step 2:

[1022] The server collects related product reviews from major e-commerce sites based on the search query received from the user. The server uses an API to retrieve reviews that match the keywords and stores them in an internal database. For example, it collects reviews related to "smartphone cases."

[1023] Step 3:

[1024] The server applies natural language processing (NLP) techniques to the collected product reviews to analyze them, including:

[1025] Review classification: Classify collected reviews as positive, negative, or neutral. For example, "The case is durable" is classified as positive, and "The price is high" is classified as negative.

[1026] Key phrase extraction: Extract key information and evaluation points from each review, such as "good design" or "durable and durable."

[1027] Sentiment analysis: Analyzing the emotional tone of a review to determine its strength, for example, distinguishing between "Very Satisfied" and "Slightly Dissatisfied."

[1028] Step 4:

[1029] The server evaluates the credibility of the reviews based on the analysis results and filters out stealth marketing and fake reviews. It does the following:

[1030] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with abnormally high ratings.

[1031] Trustworthiness Scoring: Each review is assigned a trustworthiness score to filter out fraudulent reviews. For example, reviews posted in large numbers in a short period of time are filtered out as low scores.

[1032] Step 5:

[1033] The server generates easy-to-read summaries from reviews that have passed the reliability assessment. The server succinctly summarizes key opinions and evaluation points so that users can understand them quickly. For example, it generates a summary such as, "This smartphone case is highly durable and has a well-received design, but the price is a little high."

[1034] Step 6:

[1035] The server generates personalized reviews based on the user's browsing and purchasing history. This analyzes the user's past preferences and reactions and provides optimized review information. For example, a user who prioritizes durability might be provided with a summary such as "This case is particularly durable."

[1036] Step 7:

[1037] The server uses an emotion engine to recognize the user's emotions, which involves the following steps:

[1038] Facial recognition technology: Analyzes the user's facial expressions to identify their current emotional state, for example, detecting dissatisfied facial expressions.

[1039] Voice analysis: Analyzing a user's voice to identify their emotional state, for example, determining their emotions based on their tone of voice and the way they speak.

[1040] Step 8:

[1041] The server customizes summaries based on the emotional state obtained from the emotion engine, providing users with review information that matches their emotional state. For example, if the user is in a positive mood, positive reviews will be highlighted.

[1042] Step 9:

[1043] The device displays the generated summary to the user. It is presented visually on the screen of a smartphone or computer. For example, it may display information such as, "Many users rate this case highly for its durability and design, and although it is somewhat expensive, satisfaction is high."

[1044] This is the specific processing flow of the "AI Review Summarizer" system, which allows users to quickly understand reviews and obtain information customized to their sentiment.

[1045] Example 2

[1046] 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."

[1047] Conventional product review management systems on e-commerce sites make it difficult for users to efficiently find useful information from the large number of reviews. Furthermore, stealth marketing and false reviews can make it difficult for users to make accurate decisions. Furthermore, the lack of functionality to customize reviews based on users' browsing history and emotional state makes it difficult to improve the user's purchasing experience.

[1048] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes: means for collecting product reviews based on a search query received from a user; means for analyzing the collected product reviews using natural language processing technology, categorizing the reviews, and extracting key phrases; means for evaluating the credibility of the reviews and filtering out fraudulent reviews; means for generating easy-to-read summaries from the analyzed reviews; means for generating personalized reviews based on the user's browsing history and purchasing history; means including an emotion engine for recognizing the user's emotions; means for customizing review display based on the emotion recognition results; and means for displaying the generated summaries on the user's device. This allows users to quickly and accurately obtain useful review information and make accurate decisions without being influenced by stealth marketing or fake reviews. Furthermore, the review display is customized based on the user's individual browsing history and emotional state, improving the purchasing experience.

[1049] A "search query" is a word or phrase that a user enters to search for specific information.

[1050] "Product reviews" are evaluations and comments written by users who have purchased a product about the quality and performance of that product.

[1051] "Natural language processing technology" is a technology that allows programs to understand, interpret, and generate human language.

[1052] "Classification" is the act of grouping collected data based on specific criteria.

[1053] "Keyphrase extraction" is the process of extracting important information and evaluation points from a text.

[1054] A "trustworthiness rating" is a rating used to determine whether the content of a review is trustworthy.

[1055] "Filtering" is the process of removing unwanted or harmful information.

[1056] "Summary generation" is the act of shortening a long piece of text and extracting and summarizing only the main points.

[1057] "Browsing history" is a record of pages and product information that a user has previously viewed on a website or app.

[1058] "Purchase history" is a record of products purchased by a user in the past.

[1059] A "personalized review" is a product review that is customized based on a user's individual tastes and preferences.

[1060] An "emotion engine" is a technology or system for recognizing and analyzing a user's emotional state.

[1061] "Customization" refers to adjustments or changes made to suit the needs and preferences of individual users.

[1062] A "terminal" is an electronic device used by a user to display information.

[1063] The present invention provides a system that enables users to quickly and easily select product reviews on an e-commerce site, and also has a function for recognizing user emotions and customizing the review display. A specific embodiment of this system will be described below.

[1064] 1. System Overview

[1065] The system performs a series of processes that allow users to search for product reviews and display them in a customized format. The system is designed on the premise that it is operated by a server, a user's terminal, and the user.

[1066] 2. Hardware and Software Configuration

[1067] Hardware

[1068] Server: A group of computers with high-performance processing power and large storage capacity. They are responsible for collecting and analyzing reviews on e-commerce sites.

[1069] Device: The electronic device on which a user views reviews, such as a smartphone, tablet, or computer, that is connected to the internet.

[1070] software

[1071] Natural Language Processing (NLP) technology: Software used to analyze reviews, specifically Google Cloud Natural Language and Microsoft Azure Text Analytics.

[1072] Emotion engine: Software for recognizing user emotions, including facial recognition and voice analysis technologies.

[1073] Database Management System: Software such as MongoDB or MySQL to store collected reviews and user history information.

[1074] 3. Data processing and calculation

[1075] Collecting reviews

[1076] The server collects relevant product reviews based on the search query received from the user using APIs of multiple e-commerce sites, and stores the collected reviews in an internal database.

[1077] Review analysis

[1078] The server applies NLP techniques to analyze the collected reviews. The analysis process includes the following steps:

[1079] Classifying reviews: Categorizing reviews into positive, negative, and neutral categories.

[1080] Keyphrase extraction: Extract important information and evaluation points.

[1081] Sentiment analysis: Analyze the emotional tone of each review.

[1082] filtering

[1083] The server detects and filters stealth marketing and fake reviews, performs credibility scoring, and weeds out fraudulent reviews.

[1084] Generate a summary

[1085] Generate easy-to-read summaries from reviews that have passed the credibility assessment, extracting only the most important points and summarizing them in short sentences.

[1086] Generate personalized reviews

[1087] Generates individually optimized review information based on a user's browsing and purchasing history. Analyzes past preferences and reactions to provide the most relevant information to the user.

[1088] emotion recognition

[1089] To recognize the user's emotions, the server uses an emotion engine that utilizes facial recognition and voice analysis technologies to identify the user's current emotional state.

[1090] Customized review display

[1091] Based on the user's emotional state, the server customizes the display of reviews: if the user is in a positive state, positive reviews are highlighted;

[1092] 4. Examples of concrete examples and prompts

[1093] Specific examples

[1094] When a user searches for "smartphone cases," the server collects relevant reviews from multiple e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries from reliable reviews. It then creates personalized reviews based on the user's browsing history and emotional state and displays them on the device.

[1095] Prompt Sentence Examples

[1096] If a user wants to get a detailed review summary using a generative AI model, they can enter the following prompt:

[1097] "Summarize your phone case review, especially highlighting any mentions of durability, design, and price."

[1098] In this way, the invention quickly and accurately selects and displays product reviews on a user's e-commerce site, improving the user's purchasing experience.

[1099] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1100] Step 1:

[1101] A user searches for product reviews.

[1102] Specific operation: A user accesses an e-commerce site, enters a product name or category in the search bar, for example, "smartphone case," and clicks the search button.

[1103] Input: A search query such as a product name or category.

[1104] Output: The search query information sent to the server.

[1105] Step 2:

[1106] The server collects product reviews.

[1107] Specific operation: Based on the search query received from the user, the server uses an API to collect relevant product reviews from multiple major e-commerce sites.

[1108] Input: Search query information submitted by the user.

[1109] Output: Associated product reviews stored in an internal database.

[1110] Step 3:

[1111] The server analyzes the collected reviews.

[1112] How it works: The server applies natural language processing (NLP) techniques to analyze the collected reviews, using Google Cloud Natural Language and Microsoft Azure Text Analytics. The analysis process includes review classification, key phrase extraction, and sentiment analysis.

[1113] Input: Collected product reviews.

[1114] Output: Classification results, extracted key phrases, and sentiment analysis results.

[1115] Step 4:

[1116] The server performs the trust rating and filtering.

[1117] What it does: Scores the trustworthiness of analyzed reviews and filters out stealth marketing and fake reviews. This includes the process of detecting unnatural patterns and assigning a trustworthiness score.

[1118] Input: Analysis results and review content.

[1119] Output: Trustworthy scored reviews, trustworthy reviews with fraudulent reviews filtered out.

[1120] Step 5:

[1121] The server generates an easy-to-read summary.

[1122] What it does: It generates easy-to-read summaries from reliable reviews, extracting specific evaluation points and summarizing them in short sentences. For example, it might summarize something like, "This smartphone case is durable and has a good design, but is a little pricey."

[1123] Enter: reliable reviews.

[1124] Output: A user-readable summary.

[1125] Step 6:

[1126] The server generates the personalized review.

[1127] What it does: Generates personalized reviews based on users' browsing and purchasing history, analyzing past preferences and reactions to provide users with the most relevant information.

[1128] Input: User browsing and purchasing history, trusted reviews.

[1129] Output: A personalized product review.

[1130] Step 7:

[1131] The server recognizes the user's emotions.

[1132] What it does: It uses an emotion engine to recognize the user's emotions. It uses facial recognition and voice analysis technologies to identify the user's current emotional state.

[1133] Input: User's facial image and voice data.

[1134] Output: The user's emotional state (positive, negative, etc.).

[1135] Step 8:

[1136] The server customizes the review display based on the emotional state.

[1137] Specific behavior: Based on the information obtained from the emotion engine, the summary is customized to provide review information that matches the user's emotional state. For example, if the user is in a positive mood, positive reviews will be highlighted.

[1138] Input: Emotional state information from the emotion engine, generated summary.

[1139] Output: Customized review display information.

[1140] Step 9:

[1141] Display the generated summary on your terminal.

[1142] Specific behavior: The generated summary and personalized review are displayed on the user's device. For example, a message like "Many users rate this case highly for its durability and design, and although it is a little pricey, they are very satisfied with it" is displayed on the smartphone screen.

[1143] Input: Customized review display information.

[1144] Output: Summary and review information displayed on the user's device.

[1145] (Application example 2)

[1146] 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."

[1147] Modern e-commerce sites make it difficult for users to quickly and accurately evaluate product reviews. They also contain a large number of stealth marketing and fake reviews, creating the risk of making purchasing decisions based on unreliable information. Furthermore, reviews are not personalized based on each user's preferences and emotional state, potentially reducing user satisfaction.

[1148] 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 collecting reviews based on a search query received from a user, means for analyzing the collected reviews using natural language processing technology, categorizing the reviews, and extracting key phrases, means for evaluating the credibility of the reviews and filtering out stealth marketing and false reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized summaries based on the user's browsing history and purchase history, means for analyzing the user's facial expressions using facial recognition technology and identifying their emotional state, means for analyzing the user's vocalizations using speech analysis technology and identifying their emotional state, means for customizing review displays based on information obtained from the emotion engine, and means for displaying the generated summaries on the user's terminal. This allows users to quickly and reliably obtain review information and enjoy review displays optimized based on their own emotional state and past purchase history.

[1149] A "search query" is a keyword or phrase that a user enters to search for a specific product or information.

[1150] A "review" is a piece of text that contains a user's evaluation or opinion of a product or service.

[1151] "Natural language processing technology" is a technology that allows computers to analyze, understand, and generate human language.

[1152] A "key phrase" is a word or phrase that is considered to be of high importance in a sentence.

[1153] "Stealth marketing" is a marketing technique that intentionally promotes products and services without consumers realizing it.

[1154] A "false review" is a review that contains information that is not true.

[1155] A "summary" is a short summary of the main content of the original text.

[1156] A "personalized summary" is a summary that is optimized based on an individual user's interests and past behavior.

[1157] "Facial recognition technology" is a technology that identifies and analyzes people's faces from images and videos.

[1158] "Voice analysis technology" is a technology that analyzes voice data and analyzes content, emotions, etc.

[1159] An "emotion engine" is a technology that analyzes the user's emotions and optimizes the system's operation based on the results.

[1160] This invention provides a system that enables users to quickly and easily select product reviews on an e-commerce site, and also has the function of recognizing the user's emotions and customizing the review display. Specific embodiments of this system are described below.

[1161] Generating a Program

[1162] This system is composed of a program that combines multiple methods. First, the server collects product reviews based on search queries received from users. To do this, the server retrieves relevant reviews from major e-commerce sites via API and stores them in an internal database.

[1163] Processing Overview

[1164] The following hardware and software are used to implement this program:

[1165] Hardware: Smartphone

[1166] Software: NLP analysis tools (e.g., Google Cloud Natural Language API), emotion analysis tools (e.g., Microsoft Azure Emotion API)

[1167] Database: NoSQL database (e.g. MongoDB)

[1168] Server: AWS Lambda (serverless computing)

[1169] 1. Collect reviews

[1170] The server collects reviews based on the search query submitted by the user. For example, when a user searches for "smartphone case," the server collects related reviews via the e-commerce site's API and stores them in a NoSQL database.

[1171] 2. NLP analysis

[1172] The server uses the Google Cloud Natural Language API to analyze the collected reviews, including classifying reviews as positive, negative, or neutral, extracting key phrases, and analyzing emotional tone, allowing it to quickly provide users with the information they need most.

[1173] 3. Trustworthiness Assessment and Filtering

[1174] The server evaluates the authenticity of the reviews and filters out stealth marketing and fake reviews. This step removes unusual patterns (such as a large number of posts from the same IP address) and reviews with low scores.

[1175] 4. Summary Generation

[1176] The server generates summaries from reliable reviews, which are short summaries of key features of the reviews to help users make a quick decision. For example, a summary might read, "This smartphone case is durable and has a good design, but is a bit pricey."

[1177] 5. Emotion recognition

[1178] The server uses the Microsoft Azure Emotion API to recognize the user's current emotional state. To do this, it uses facial expression analysis using the user's smartphone camera and emotion analysis technology from voice. This allows it to identify the user's emotional state.

[1179] 6. Personalized Display

[1180] The server customizes the display of reviews based on information obtained from the emotion engine. For example, when a user is in a positive emotional state, it highlights positive reviews. This function provides information optimized for each user.

[1181] 7. Display

[1182] Finally, the generated summary is displayed on the user's device. If the device is a smartphone, the appropriate summary is visually presented on the screen.

[1183] Specific examples

[1184] When a user searches for "smartphone case," the server sends the following prompt to the generative AI model:

[1185] Get product-related reviews and categorize them as positive, negative, or neutral based on sentiment. Extract key phrases and generate summaries. Eliminate stealth marketing patterns and assess trustworthiness. Display customized summaries based on the user's emotional state.

[1186] This allows users to quickly obtain reliable review information and enjoy optimized review displays based on their own emotional state and past purchasing history.

[1187] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1188] Step 1:

[1189] A user enters keywords related to a specific product name or category into the search bar and clicks the search button. This input generates a search query to retrieve reviews related to the product. The server receives this search query and secures the data to proceed to the next step.

[1190] Step 2:

[1191] The server collects relevant product reviews from major e-commerce sites based on the received search query. Specifically, it retrieves the relevant reviews via API and stores them in a NoSQL database. For example, it collects hundreds of reviews about "smartphone cases." This allows review information based on the product the user is looking for to be stored in the database.

[1192] Step 3:

[1193] The server applies natural language processing (NLP) technology to analyze the collected product reviews. First, it classifies the review content as positive, negative, or neutral. Next, it extracts important information and evaluation points (key phrases) from each review. It also performs sentiment analysis of the reviews to determine the emotional tone of the review. For example, it distinguishes between reviews that are "very satisfied" and reviews that are "slightly dissatisfied." This allows the information users need to be provided quickly.

[1194] Step 4:

[1195] The server evaluates the credibility of reviews based on the analysis results and filters out stealth marketing and fake reviews. First, it detects unnatural patterns and removes reviews posted in large numbers from the same IP address or reviews with abnormally high ratings. Next, it assigns a credibility score to each review and eliminates reviews with low scores. This step allows reliable reviews to be selected.

[1196] Step 5:

[1197] The server generates easy-to-read summaries from reviews that pass the reliability assessment. Specifically, it summarizes important evaluation points in a concise manner, allowing users to quickly obtain key information without having to read lengthy reviews. For example, a summary might be generated such as, "This smartphone case is highly durable and has a well-received design, but is a little pricey."

[1198] Step 6:

[1199] The server generates personalized summaries based on the user's browsing and purchasing history. It analyzes the types of products the user has previously purchased and provides summaries tailored to those preferences. For example, a user who has previously preferred durable products might receive a summary such as "This case is particularly durable."

[1200] Step 7:

[1201] The server uses an emotion engine to recognize the user's emotions. To do this, it analyzes the user's facial expressions using the device (smartphone) camera and identifies the user's emotional state using the Microsoft Azure Emotion API. It also analyzes the user's vocalizations to recognize the user's emotional state.

[1202] Step 8:

[1203] The server customizes the review display based on the information obtained from the emotion engine and the personalized summary. For example, if the user is in a positive state, it highlights positive reviews, and if the user is in an anxious state, it highlights reliable reviews. This step provides the user with the most appropriate information according to their emotional state.

[1204] Step 9:

[1205] The generated summary is displayed on the device (smartphone) screen. Users can quickly obtain the appropriate summary information to quickly evaluate the product. For example, a summary such as "Many users rate this case highly for its durability and design, and although the price is somewhat high, satisfaction is high" may be displayed.

[1206] 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.

[1207] 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.

[1208] 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.

[1209] [Fourth embodiment]

[1210] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1211] 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.

[1212] 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).

[1213] 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.

[1214] 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.

[1215] 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).

[1216] 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.

[1217] 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.

[1218] 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.

[1219] 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.

[1220] 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.

[1221] 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.

[1222] 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."

[1223] The present invention provides a system that allows users to quickly and easily select product reviews on an e-commerce site. This system is implemented using the following programs in particular.

[1224] Program processing explanation

[1225] 1. User searches for product reviews

[1226] A user visits an e-commerce site and enters keywords related to a specific product name or category into the search bar. Based on this search query, the server begins searching for relevant reviews. For example, a user enters "smartphone case" and clicks the search button.

[1227] 2. The server collects reviews

[1228] The server collects relevant product reviews from major e-commerce sites based on the search query entered by the user. The reviews are retrieved via API and stored in an internal database. This results in the collection of hundreds of reviews related to "smartphone cases."

[1229] 3. Review Analysis Using Natural Language Processing (NLP)

[1230] The server analyzes the collected reviews using natural language processing technology, which performs the following processes:

[1231] Review categorization: Categorize reviews as positive, negative, or neutral. For example, based on the content of the review, the part that says "high durability" can be categorized as positive, and conversely, the part that says "high price" can be categorized as negative.

[1232] Key phrase extraction: Extracting important information from each review, such as "nice design" or "durable and durable."

[1233] Sentiment analysis: Analyzing the tone of a review to determine emotional intensity, for example, distinguishing between "very satisfied" and "somewhat dissatisfied" reviews.

[1234] 4. Filtering and Eliminating Stealth Marketing

[1235] The server then uses the analysis results to assess trustworthiness and filter out fraudulent reviews, including:

[1236] Unnatural pattern detection: Automatically detects reviews posted in large numbers from the same IP address or reviews with abnormally high ratings.

[1237] Trustworthiness scoring: Each review is assigned a certain score to filter out fraudulent reviews. For example, reviews submitted in large numbers in a short period of time are given a low score and are subject to elimination.

[1238] 5. Generating a Summary

[1239] The server generates an easy-to-read summary based on the filtered reviews. This allows users to quickly obtain only the main points of evaluation without having to read lengthy reviews. For example, it generates a summary such as, "This smartphone case is highly durable and has a well-received design, but is a little pricey."

[1240] 6. Providing personalized reviews

[1241] The server generates personalized reviews based on the user's browsing and purchasing history. For example, if a user has previously purchased products that emphasize durability, the server provides a summary that highlights information such as, "This case is particularly durable."

[1242] 7. Display summary on terminal

[1243] The generated summary is displayed on the user's device. Summary information about the product is displayed on the screen of a smartphone or PC, allowing the user to quickly check important information. For example, a message may appear on a smartphone screen saying, "Many users rate this case highly for its durability and design, and although the price is somewhat high, satisfaction is high."

[1244] Specific examples

[1245] When a user searches for "smartphone case" on their smartphone, the server collects relevant reviews from e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries. These summaries are then displayed as personalized summaries on the user's device, allowing the user to quickly and efficiently evaluate the product.

[1246] A system based on this claim would speed up understanding of product reviews, increase their reliability, and improve the consumer buying experience.

[1247] The processing flow will be explained below.

[1248] Step 1:

[1249] A user types keywords related to a specific product name or category into the search bar of an e-commerce site, and is ready to find reviews that match their interests.

[1250] Step 2:

[1251] The server collects relevant product reviews from major e-commerce sites based on the search query received from the user. The server uses an API to retrieve reviews that match the keywords and stores them in an internal database.

[1252] Step 3:

[1253] The server starts analyzing the collected product reviews by applying natural language processing (NLP) techniques. The server performs the following processes:

[1254] Classification of reviews: Positive, negative, or neutral.

[1255] Key phrase extraction: Extract important information and evaluation points.

[1256] Sentiment analysis: Analyze the emotional tone of each review.

[1257] Step 4:

[1258] The server uses the analysis results to assess the trustworthiness of reviews and filter out stealth marketing and fake reviews. This includes:

[1259] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with only unusually high ratings.

[1260] Trustworthiness Scoring: Each review is assigned a trustworthiness score to help weed out fraudulent reviews.

[1261] Step 5:

[1262] The server generates easy-to-read summaries from reviews that have passed the credibility assessment. The server concisely extracts key opinions and evaluation points and summarizes them in a format that users can quickly understand.

[1263] Step 6:

[1264] The server generates personalized reviews based on the user's browsing and purchasing history, analyzing the user's past preferences and reactions and providing individually optimized review information.

[1265] Step 7:

[1266] The terminal displays the generated summary information to the user, who can then visually view the generated summary on the screen of their smartphone or computer, allowing them to quickly check the product evaluation.

[1267] This is the specific processing flow of the "AI Review Summarizer" system, which enables users to quickly organize and accurately understand vast amounts of review information.

[1268] Example 1

[1269] 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."

[1270] When purchasing a product on an online e-commerce site, users need to find useful information from numerous reviews, but this task requires time and effort. Furthermore, some reviews may be fraudulent or unreliable, which can lead to misjudgments. Furthermore, because different users have different priorities, it is not always easy to find the most suitable review for each user. Therefore, there is a need for a fast and reliable review selection system that can provide personalized reviews tailored to users' preferences.

[1271] 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.

[1272] In this invention, the server includes means for collecting product reviews based on a search query received from a user, means for analyzing the collected product reviews using natural language processing technology, categorizing the reviews, and extracting key phrases, means for evaluating the reliability of the reviews and filtering out fraudulent reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized summaries based on the user's browsing history and purchase history, and means for providing the generated summaries to a display device. This allows users to quickly and efficiently obtain reliable product reviews and make purchasing decisions based on information that suits their needs.

[1273] A "search query" is a keyword or phrase that a user enters to search for information related to a particular product or category.

[1274] A "product review" is a written description of the product's usage and evaluation by a user who has purchased the product on an e-commerce site or the like.

[1275] "Natural language processing technology" is a technology that enables computers to understand and analyze human language and extract its meaning.

[1276] "Review categorization" is the process of dividing collected product reviews into categories based on their evaluation, such as positive, negative, or neutral.

[1277] A "key phrase" is a major phrase in a piece of text that expresses important information or a particular theme.

[1278] "Review credibility assessment" is the process of determining how accurate and trustworthy the collected reviews are.

[1279] "Fraudulent reviews" are false or exaggerated reviews posted with the intent to manipulate product sales.

[1280] "Filtering" is the process of removing unnecessary or fraudulent information from collected data.

[1281] A "summary" is a concise summary of the main points extracted from a lengthy review.

[1282] A "personalized summary" is a summary that is individually customized based on a particular user's preferences and purchasing history.

[1283] A "display device" is a device for visually presenting the generated summary or other information to a user, including a smartphone or personal computer.

[1284] This invention provides an information processing system that allows users to quickly and easily select product reviews on e-commerce sites. The system collects relevant reviews based on a user's search query, analyzes them using natural language processing (NLP), evaluates their credibility, generates summaries, and provides personalized summaries to users.

[1285] System Configuration

[1286] The system consists of the following main components:

[1287] 1. User device: The device where a user enters a search query and sees the results. This can be a smartphone or a computer.

[1288] 2. Server: A central processing unit that receives search queries from users and collects, analyzes, filters, and summarizes reviews. The server is made up of hardware with a high-performance processor and large memory capacity.

[1289] 3. Natural Language Processing Engine: Software for analyzing reviews. NLP tools such as Python's NLTK (Natural Language Toolkit) and Gensim are used.

[1290] Data processing and calculation

[1291] The server does the following:

[1292] 1. Review collection: The server collects relevant product reviews using the APIs of major e-commerce sites based on the search query received from the user. For example, if a user searches for "smartphone case," the server retrieves data via the APIs of Amazon and Rakuten Ichiba.

[1293] 2. Review Analysis: We analyze the collected reviews using natural language processing (NLP). We classify reviews as "positive," "negative," or "neutral," extract key phrases, and perform sentiment analysis. We use the NLTK and Gensim libraries to perform this analysis.

[1294] 3. Review credibility evaluation and filtering: The server evaluates the credibility of reviews based on the analysis results and filters out fraudulent reviews. The algorithms used include unnatural pattern detection and credibility scoring.

[1295] 4. Summary Generation: The server generates summaries from the filtered reviews. It also creates personalized summaries that emphasize different points for each user. It uses the Gensim library to generate the summary text.

[1296] Specific examples

[1297] For example:

[1298] When a user searches for "smartphone case" on their smartphone, the following prompt is entered into the system:

[1299] "I want to generate summaries from reviews about smartphone cases, classify them as positive or negative, and extract key phrases."

[1300] The server receives the search query and collects review information via APIs from Amazon and Rakuten. It then analyzes the reviews using NLTK and Gensim and classifies them as positive, negative, or neutral. It then scores the reviews' trustworthiness, filters out unnatural patterns, and generates a personalized summary tailored to the user's preferences. This summary is then displayed on the browser screen of a smartphone or PC, allowing users to easily check important information about the review.

[1301] This invention enables users to quickly and efficiently obtain highly reliable product reviews and make purchasing decisions based on information that meets their needs.

[1302] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1303] Step 1:

[1304] User searches for product reviews

[1305] A user accesses an e-commerce site using a smartphone or PC and enters keywords related to a specific product name or category into the search bar. For example, if the user enters "smartphone case," the search query is sent to the server when the user clicks the search button.

[1306] Input: The search query entered by the user (e.g., "smartphone case")

[1307] Output: The search query sent to the server

[1308] Step 2:

[1309] Server collects reviews

[1310] The server uses the APIs of major e-commerce sites to collect relevant product reviews based on the search query received from the user. Secure data acquisition is performed using an API key, and the acquired review information is stored in an internal database. For example, reviews on "smartphone cases" are collected using the APIs of Amazon and Rakuten Ichiba.

[1311] Input: The search query received by the server

[1312] Output: A database of collected review information

[1313] Step 3:

[1314] Review analysis using natural language processing (NLP)

[1315] The server analyzes the collected reviews using natural language processing (NLP) technology. Specifically, it performs the following processes:

[1316] Review Classification: Classify review text as positive, negative, or neutral. For example, "The case is very sturdy" is classified as positive, and "The price is too high" is classified as negative.

[1317] Keyphrase extraction: Important keyphrases are extracted using TF-IDF and Word2Vec algorithms. Examples include "good design" and "durable."

[1318] Sentiment analysis: Analyzes the sentiment of the review and evaluates the strength of the sentiment. For example, a "Very Satisfied" review will have a high positive score, while a "Slightly Dissatisfied" review will have a negative score.

[1319] Input: Collected review information

[1320] Output: Analyzed review classification, key phrases, and sentiment analysis data

[1321] Step 4:

[1322] Review credibility rating and filtering

[1323] The server evaluates the reliability of reviews based on the analysis results and performs filtering. Specifically, it performs the following processes:

[1324] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with only unusually high ratings.

[1325] Trustworthiness Scoring: Assign a score to each review and automatically filter out fraudulent reviews with low scores.

[1326] Input: Parsed review classification, key phrases, and sentiment analysis data

[1327] Output: Filtered data containing only reliable reviews

[1328] Step 5:

[1329] Generate a summary

[1330] The server generates summaries based on the filtered reviews. Using the Gensim library, it extracts key points from the review text and summarizes them in a concise format to create easy-to-read summaries. For example, it generates a summary such as, "This smartphone case is durable and has a good design, but is a little pricey."

[1331] Input: Filtered reviews

[1332] Output: Generated summary

[1333] Step 6:

[1334] Generate personalized reviews

[1335] The server generates a personalized summary based on the user's browsing and purchasing history. For a user who has previously purchased products that emphasize durability, the server generates a summary that emphasizes the fact that "this case is particularly durable."

[1336] Input: User browsing history, purchase history, filtered reviews

[1337] Output: Personalized Summary

[1338] Step 7:

[1339] Display summary to terminal

[1340] The generated summary is displayed on the user's device. Summary information about the product is displayed on the screen of a smartphone or computer, allowing the user to quickly check important information. Specifically, the smartphone browser or dedicated app screen displays the message, "Many users rate this case highly for its durability and design, and although the price is a little high, satisfaction is high."

[1341] Input: Generated summary, personalized summary

[1342] Output: Summary information displayed on the user's terminal

[1343] (Application example 1)

[1344] 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."

[1345] Online e-commerce sites contain a huge number of product reviews, making it difficult for users to quickly access the information they need. Conventional systems lack the ability to evaluate and filter the reliability of reviews, and many contain stealth marketing and fraudulent reviews. Furthermore, personalized reviews tailored to each user's needs are not provided, making it difficult for users to effectively utilize reviews. This can lead to consumers making incorrect decisions or wasting unnecessary time when choosing products.

[1346] 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.

[1347] In this invention, the server includes means for collecting product reviews based on a search query received from a user, means for analyzing the collected product reviews using natural language processing technology to categorize the reviews and extract key phrases, means for evaluating the reliability of the reviews and filtering out stealth marketing and false reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized reviews based on the user's browsing history and purchasing history, means for displaying the generated summaries on the user's device, means for analyzing the content of the reviews and categorizing them as positive, negative, or neutral, and means for analyzing detailed information about the products viewed by the user and generating prompt sentences. This allows users to quickly access reliable review summaries and efficiently obtain information tailored to their purchasing behavior.

[1348] A "search query" is a keyword or phrase that a user enters when searching for a product or information.

[1349] A "product review" is a record of the evaluation and impressions of a product made by a user who has purchased the product.

[1350] "Natural language processing technology" is a technology that allows computers to understand, analyze, and generate human language.

[1351] "Keyphrases" refer to important words or phrases within a review or document.

[1352] "Stealth marketing" refers to advertising activities that are carried out without consumers realizing that they are advertising.

[1353] A "false review" is a review that contains false information that differs from the actual performance or usability of the product.

[1354] A "summary" is a short summary that extracts the main points from a long document or review.

[1355] A "personalized review" is a review that is customized based on the interests and preferences of an individual user.

[1356] "Evaluating credibility" refers to the act of judging and evaluating the accuracy and reliability of collected reviews.

[1357] This invention provides a system that allows users to quickly and easily select product reviews on an e-commerce site. This system is realized through the following steps.

[1358] First, a user accesses an e-commerce site from a smartphone or computer and enters a search query related to a specific product name or category. This search query is sent to a server, which then collects relevant product reviews from the e-commerce site. The collected reviews are then stored in the server's database.

[1359] The server then analyzes the collected reviews using natural language processing (NLP) technology. This analysis includes categorizing reviews into positive, negative, and neutral categories and extracting key phrases. Specific analysis techniques used include tools such as the Python nltk library and scikit-learn.

[1360] The server then evaluates the credibility of reviews based on the analysis results and filters out fraudulent reviews. This evaluation is based on review posting frequency and IP address pattern analysis. It also applies algorithms to automatically detect stealth marketing and fake reviews. This ensures users have access to reliable review information.

[1361] The server generates summaries of the analyzed and filtered reviews. These summaries are short summaries that extract the main points of evaluation from the long reviews. For example, a summary might be generated that reads, "This smartphone case is durable and has a good design, but is a little pricey."

[1362] The server also generates personalized reviews based on the user's browsing and purchasing history. These personalized reviews provide information customized to the user's preferences and interests. For example, for a user who has previously purchased a durable product, the server may emphasize that "this case is particularly durable."

[1363] Finally, the generated summary and personalized review are displayed on the user's device. Users can view the concise review information on their smartphone or computer screen, allowing them to quickly and efficiently understand product ratings and obtain information that will help them make a purchase.

[1364] For example, when a user searches for "smartphone case" on their smartphone, the server collects relevant reviews from e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates a summary, which is then displayed as a personalized summary on the user's device.

[1365] An example prompt is, "Based on the product name searched by the user, collect relevant reviews, analyze the reviews using natural language processing, and generate summaries. In particular, display them separately from the perspectives of positive, negative, and neutral."

[1366] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1367] Step 1:

[1368] A user accesses an e-commerce site using a smartphone or computer and enters a search query related to a specific product name or category. The search query is sent from the device to the server. The input data at this time is the keywords or phrases entered by the user.

[1369] Step 2:

[1370] The server collects relevant product reviews from major e-commerce sites based on the received search query. The collection is done using an API, and the acquired review information is stored in the server's database. The output data at this time is raw review information.

[1371] Step 3:

[1372] The server uses natural language processing (NLP) technology to analyze the collected product reviews. Specifically, it categorizes the reviews into positive, negative, and neutral categories and extracts important key phrases. The NLP technology used includes Python's nltk library and scikit-learn. The input data is raw review information, and the output data is analyzed review information.

[1373] Step 4:

[1374] The server evaluates the reliability of the reviews based on the analysis results and filters out fraudulent reviews. The filtering method uses an algorithm that automatically detects large numbers of posts from the same IP address and reviews with abnormally high ratings. The input data is the analyzed review information, and the output data is the reliability-evaluated review information.

[1375] Step 5:

[1376] The server generates easy-to-read summaries from the reviews that have been evaluated for trustworthiness. The summary generation uses technology to extract the main evaluation points that have been analyzed and summarize them in a short form. For example, a summary such as "It is highly durable and has a popular design, but it is expensive" may be generated. The input data at this time is the review information that has been evaluated for trustworthiness, and the output data is the summary information.

[1377] Step 6:

[1378] The server generates personalized reviews based on the user's browsing history and purchase history. This utilizes information such as the user's past product preferences. For example, for a user who prioritizes durability, the server emphasizes information such as "This case is particularly durable." The input data at this time are summary information and user history information, and the output data is personalized summary information.

[1379] Step 7:

[1380] The generated summary and personalized review are displayed on the user's device. The user can view the concisely summarized review information on the screen of a smartphone or computer. The input data is the personalized summary information, and the output data is the summary information displayed on the user's device.

[1381] Step 8:

[1382] The server analyzes detailed information about the product viewed by the user and generates a prompt sentence. This prompt sentence is used to complement the user's next search query and review confirmation. For example, a prompt sentence may be generated that reads, "Based on the product name searched by the user, please collect related reviews, analyze the reviews using natural language processing, and generate summaries. Please display them in particular from the perspectives of positive, negative, and neutral." The input data at this time are browsing history information and review information, and the output data is the prompt sentence.

[1383] 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.

[1384] The present invention is a system that enables users to quickly and easily select product reviews on e-commerce sites, and also has the ability to recognize users' emotions and customize the review display. The specific program processing will be explained in detail below, with examples.

[1385] Program processing explanation

[1386] 1. User searches for product reviews

[1387] A user visits an e-commerce site and enters keywords related to a specific product name or category into the search bar, which prepares the site to retrieve reviews tailored to the user's interests. For example, a user enters "smartphone cases" and clicks the search button.

[1388] 2. The server collects reviews

[1389] The server collects relevant product reviews from major e-commerce sites based on the search query received from the user. The reviews are retrieved via API and stored in an internal database. For example, hundreds of reviews on "smartphone cases" are collected.

[1390] 3. Review Analysis Using Natural Language Processing (NLP)

[1391] The server analyzes the collected product reviews using natural language processing (NLP) techniques. It performs the following processes:

[1392] Review categorization: Categorize reviews as positive, negative, or neutral. For example, categorize reviews that say "high durability" as positive and reviews that say "high price" as negative.

[1393] Key phrase extraction: Extract important information and evaluation points, such as "good design" or "durable and durable."

[1394] Sentiment analysis: Analyzing the emotional tone of each review, for example, distinguishing between "Very Satisfied" and "Somewhat Dissatisfied."

[1395] 4. Filtering and Eliminating Stealth Marketing

[1396] The server evaluates the credibility of the reviews based on the analysis results and filters out stealth marketing and fake reviews. This includes the following actions:

[1397] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with an unusually high number of positive ratings.

[1398] Credibility scoring: Each review is assigned a credibility score to filter out fraudulent reviews. For example, reviews posted in large numbers in a short period of time will be assigned a low score and will be subject to elimination.

[1399] 5. Generating a Summary

[1400] The server generates easy-to-read summaries from reviews that pass the trust assessment. Users can quickly obtain only the main points of evaluation without having to read lengthy reviews. For example, a summary such as "This smartphone case is highly durable and has a well-received design, but is a little pricey" can be generated.

[1401] 6. Providing personalized reviews

[1402] The server generates personalized reviews based on the user's browsing and purchasing history. It analyzes the user's past preferences and reactions and provides individually optimized review information. In particular, for users who have previously preferred products that emphasize durability, it provides a summary such as "This case is particularly durable."

[1403] 7. Emotion Recognition with Emotion Engine

[1404] The server uses an emotion engine to recognize the user's emotion. It performs the following operations:

[1405] Facial recognition technology: Analyzes the user's facial expressions to identify their current emotional state. For example, it can detect when a user is unhappy.

[1406] Voice analysis: Analyzing the user's vocal utterances to identify their emotional state, for example, by analyzing the user's tone of voice and word choice.

[1407] 8. Customize review display based on sentiment

[1408] The server customizes the summary based on the information obtained from the emotion engine, providing review information that matches the user's emotional state. For example, if the user is in a positive mood, the server highlights the same positive reviews.

[1409] 9. Display summary on terminal

[1410] The generated summary is displayed on the user's device. The device visually presents the generated summary on the screen of a smartphone or PC. For example, a smartphone screen might say, "Many users rate this case highly for its durability and design, and although the price is a little high, satisfaction is high."

[1411] Specific examples

[1412] When a user searches for "smartphone case," the server collects related reviews and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries from reliable reviews. It then creates personalized reviews based on the user's browsing history and emotional state and displays them on the device. This allows users to quickly and accurately evaluate products.

[1413] The system according to this claim enhances the consumer buying experience by speeding up the understanding of product reviews and customizing the display according to the user's sentiment.

[1414] The processing flow will be explained below.

[1415] Step 1:

[1416] A user visits an e-commerce site and types keywords related to a specific product name or category into the search bar, generating a search query to find reviews that match the user's interests.

[1417] Step 2:

[1418] The server collects related product reviews from major e-commerce sites based on the search query received from the user. The server uses an API to retrieve reviews that match the keywords and stores them in an internal database. For example, it collects reviews related to "smartphone cases."

[1419] Step 3:

[1420] The server applies natural language processing (NLP) techniques to the collected product reviews to analyze them, including:

[1421] Review classification: Classify collected reviews as positive, negative, or neutral. For example, "The case is durable" is classified as positive, and "The price is high" is classified as negative.

[1422] Key phrase extraction: Extract key information and evaluation points from each review, such as "good design" or "durable and durable."

[1423] Sentiment analysis: Analyzing the emotional tone of a review to determine its strength, for example, distinguishing between "Very Satisfied" and "Slightly Dissatisfied."

[1424] Step 4:

[1425] The server evaluates the credibility of the reviews based on the analysis results and filters out stealth marketing and fake reviews. It does the following:

[1426] Unnatural pattern detection: Detects reviews posted in large numbers from the same IP address or reviews with abnormally high ratings.

[1427] Trustworthiness Scoring: Each review is assigned a trustworthiness score to filter out fraudulent reviews. For example, reviews posted in large numbers in a short period of time are filtered out as low scores.

[1428] Step 5:

[1429] The server generates easy-to-read summaries from reviews that have passed the reliability assessment. The server succinctly summarizes key opinions and evaluation points so that users can understand them quickly. For example, it generates a summary such as, "This smartphone case is highly durable and has a well-received design, but the price is a little high."

[1430] Step 6:

[1431] The server generates personalized reviews based on the user's browsing and purchasing history. This analyzes the user's past preferences and reactions and provides optimized review information. For example, a user who prioritizes durability might be provided with a summary such as "This case is particularly durable."

[1432] Step 7:

[1433] The server uses an emotion engine to recognize the user's emotions, which involves the following steps:

[1434] Facial recognition technology: Analyzes the user's facial expressions to identify their current emotional state, for example, detecting dissatisfied facial expressions.

[1435] Voice analysis: Analyzing a user's voice to identify their emotional state, for example, determining their emotions based on their tone of voice and the way they speak.

[1436] Step 8:

[1437] The server customizes summaries based on the emotional state obtained from the emotion engine, providing users with review information that matches their emotional state. For example, if the user is in a positive mood, positive reviews will be highlighted.

[1438] Step 9:

[1439] The device displays the generated summary to the user. It is presented visually on the screen of a smartphone or computer. For example, it may display information such as, "Many users rate this case highly for its durability and design, and although it is somewhat expensive, satisfaction is high."

[1440] This is the specific processing flow of the "AI Review Summarizer" system, which allows users to quickly understand reviews and obtain information customized to their sentiment.

[1441] Example 2

[1442] 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."

[1443] Conventional product review management systems on e-commerce sites make it difficult for users to efficiently find useful information from the large number of reviews. Furthermore, stealth marketing and false reviews can make it difficult for users to make accurate decisions. Furthermore, the lack of functionality to customize reviews based on users' browsing history and emotional state makes it difficult to improve the user's purchasing experience.

[1444] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes: means for collecting product reviews based on a search query received from a user; means for analyzing the collected product reviews using natural language processing technology, categorizing the reviews, and extracting key phrases; means for evaluating the credibility of the reviews and filtering out fraudulent reviews; means for generating easy-to-read summaries from the analyzed reviews; means for generating personalized reviews based on the user's browsing history and purchasing history; means including an emotion engine for recognizing the user's emotions; means for customizing review display based on the emotion recognition results; and means for displaying the generated summaries on the user's device. This allows users to quickly and accurately obtain useful review information and make accurate decisions without being influenced by stealth marketing or fake reviews. Furthermore, the review display is customized based on the user's individual browsing history and emotional state, improving the purchasing experience.

[1445] A "search query" is a word or phrase that a user enters to search for specific information.

[1446] "Product reviews" are evaluations and comments written by users who have purchased a product about the quality and performance of that product.

[1447] "Natural language processing technology" is a technology that allows programs to understand, interpret, and generate human language.

[1448] "Classification" is the act of grouping collected data based on specific criteria.

[1449] "Keyphrase extraction" is the process of extracting important information and evaluation points from a text.

[1450] A "trustworthiness rating" is a rating used to determine whether the content of a review is trustworthy.

[1451] "Filtering" is the process of removing unwanted or harmful information.

[1452] "Summary generation" is the act of shortening a long piece of text and extracting and summarizing only the main points.

[1453] "Browsing history" is a record of pages and product information that a user has previously viewed on a website or app.

[1454] "Purchase history" is a record of products purchased by a user in the past.

[1455] A "personalized review" is a product review that is customized based on a user's individual tastes and preferences.

[1456] An "emotion engine" is a technology or system for recognizing and analyzing a user's emotional state.

[1457] "Customization" refers to adjustments or changes made to suit the needs and preferences of individual users.

[1458] A "terminal" is an electronic device used by a user to display information.

[1459] The present invention provides a system that enables users to quickly and easily select product reviews on an e-commerce site, and also has a function for recognizing user emotions and customizing the review display. A specific embodiment of this system will be described below.

[1460] 1. System Overview

[1461] The system performs a series of processes that allow users to search for product reviews and display them in a customized format. The system is designed on the premise that it is operated by a server, a user's terminal, and the user.

[1462] 2. Hardware and Software Configuration

[1463] Hardware

[1464] Server: A group of computers with high-performance processing power and large storage capacity. They are responsible for collecting and analyzing reviews on e-commerce sites.

[1465] Device: The electronic device on which a user views reviews, such as a smartphone, tablet, or computer, that is connected to the internet.

[1466] software

[1467] Natural Language Processing (NLP) technology: Software used to analyze reviews, specifically Google Cloud Natural Language and Microsoft Azure Text Analytics.

[1468] Emotion engine: Software for recognizing user emotions, including facial recognition and voice analysis technologies.

[1469] Database Management System: Software such as MongoDB or MySQL to store collected reviews and user history information.

[1470] 3. Data processing and calculation

[1471] Collecting reviews

[1472] The server collects relevant product reviews based on the search query received from the user using APIs of multiple e-commerce sites, and stores the collected reviews in an internal database.

[1473] Review analysis

[1474] The server applies NLP techniques to analyze the collected reviews. The analysis process includes the following steps:

[1475] Classifying reviews: Categorizing reviews into positive, negative, and neutral categories.

[1476] Keyphrase extraction: Extract important information and evaluation points.

[1477] Sentiment analysis: Analyze the emotional tone of each review.

[1478] filtering

[1479] The server detects and filters stealth marketing and fake reviews, performs credibility scoring, and weeds out fraudulent reviews.

[1480] Generate a summary

[1481] Generate easy-to-read summaries from reviews that have passed the credibility assessment, extracting only the most important points and summarizing them in short sentences.

[1482] Generate personalized reviews

[1483] Generates individually optimized review information based on a user's browsing and purchasing history. Analyzes past preferences and reactions to provide the most relevant information to the user.

[1484] emotion recognition

[1485] To recognize the user's emotions, the server uses an emotion engine that utilizes facial recognition and voice analysis technologies to identify the user's current emotional state.

[1486] Customized review display

[1487] Based on the user's emotional state, the server customizes the display of reviews: if the user is in a positive state, positive reviews are highlighted;

[1488] 4. Examples of concrete examples and prompts

[1489] Specific examples

[1490] When a user searches for "smartphone cases," the server collects relevant reviews from multiple e-commerce sites and analyzes them using NLP technology. It then filters out fraudulent reviews and generates summaries from reliable reviews. It then creates personalized reviews based on the user's browsing history and emotional state and displays them on the device.

[1491] Prompt Sentence Examples

[1492] If a user wants to get a detailed review summary using a generative AI model, they can enter the following prompt:

[1493] "Summarize your phone case review, especially highlighting any mentions of durability, design, and price."

[1494] In this way, the invention quickly and accurately selects and displays product reviews on a user's e-commerce site, improving the user's purchasing experience.

[1495] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1496] Step 1:

[1497] A user searches for product reviews.

[1498] Specific operation: A user accesses an e-commerce site, enters a product name or category in the search bar, for example, "smartphone case," and clicks the search button.

[1499] Input: A search query such as a product name or category.

[1500] Output: The search query information sent to the server.

[1501] Step 2:

[1502] The server collects product reviews.

[1503] Specific operation: Based on the search query received from the user, the server uses an API to collect relevant product reviews from multiple major e-commerce sites.

[1504] Input: Search query information submitted by the user.

[1505] Output: Associated product reviews stored in an internal database.

[1506] Step 3:

[1507] The server analyzes the collected reviews.

[1508] How it works: The server applies natural language processing (NLP) techniques to analyze the collected reviews, using Google Cloud Natural Language and Microsoft Azure Text Analytics. The analysis process includes review classification, key phrase extraction, and sentiment analysis.

[1509] Input: Collected product reviews.

[1510] Output: Classification results, extracted key phrases, and sentiment analysis results.

[1511] Step 4:

[1512] The server performs the trust rating and filtering.

[1513] What it does: Scores the trustworthiness of analyzed reviews and filters out stealth marketing and fake reviews. This includes the process of detecting unnatural patterns and assigning a trustworthiness score.

[1514] Input: Analysis results and review content.

[1515] Output: Trustworthy scored reviews, trustworthy reviews with fraudulent reviews filtered out.

[1516] Step 5:

[1517] The server generates an easy-to-read summary.

[1518] What it does: It generates easy-to-read summaries from reliable reviews, extracting specific evaluation points and summarizing them in short sentences. For example, it might summarize something like, "This smartphone case is durable and has a good design, but is a little pricey."

[1519] Enter: reliable reviews.

[1520] Output: A user-readable summary.

[1521] Step 6:

[1522] The server generates the personalized review.

[1523] What it does: Generates personalized reviews based on users' browsing and purchasing history, analyzing past preferences and reactions to provide users with the most relevant information.

[1524] Input: User browsing and purchasing history, trusted reviews.

[1525] Output: A personalized product review.

[1526] Step 7:

[1527] The server recognizes the user's emotions.

[1528] What it does: It uses an emotion engine to recognize the user's emotions. It uses facial recognition and voice analysis technologies to identify the user's current emotional state.

[1529] Input: User's facial image and voice data.

[1530] Output: The user's emotional state (positive, negative, etc.).

[1531] Step 8:

[1532] The server customizes the review display based on the emotional state.

[1533] Specific behavior: Based on the information obtained from the emotion engine, the summary is customized to provide review information that matches the user's emotional state. For example, if the user is in a positive mood, positive reviews will be highlighted.

[1534] Input: Emotional state information from the emotion engine, generated summary.

[1535] Output: Customized review display information.

[1536] Step 9:

[1537] Display the generated summary on your terminal.

[1538] Specific behavior: The generated summary and personalized review are displayed on the user's device. For example, a message like "Many users rate this case highly for its durability and design, and although it is a little pricey, they are very satisfied with it" is displayed on the smartphone screen.

[1539] Input: Customized review display information.

[1540] Output: Summary and review information displayed on the user's device.

[1541] (Application example 2)

[1542] 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."

[1543] Modern e-commerce sites make it difficult for users to quickly and accurately evaluate product reviews. They also contain a large number of stealth marketing and fake reviews, creating the risk of making purchasing decisions based on unreliable information. Furthermore, reviews are not personalized based on each user's preferences and emotional state, potentially reducing user satisfaction.

[1544] 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 collecting reviews based on a search query received from a user, means for analyzing the collected reviews using natural language processing technology, categorizing the reviews, and extracting key phrases, means for evaluating the credibility of the reviews and filtering out stealth marketing and false reviews, means for generating easy-to-read summaries from the analyzed reviews, means for generating personalized summaries based on the user's browsing history and purchase history, means for analyzing the user's facial expressions using facial recognition technology and identifying their emotional state, means for analyzing the user's vocalizations using speech analysis technology and identifying their emotional state, means for customizing review displays based on information obtained from the emotion engine, and means for displaying the generated summaries on the user's terminal. This allows users to quickly and reliably obtain review information and enjoy review displays optimized based on their own emotional state and past purchase history.

[1545] A "search query" is a keyword or phrase that a user enters to search for a specific product or information.

[1546] A "review" is a piece of text that contains a user's evaluation or opinion of a product or service.

[1547] "Natural language processing technology" is a technology that allows computers to analyze, understand, and generate human language.

[1548] A "key phrase" is a word or phrase that is considered to be of high importance in a sentence.

[1549] "Stealth marketing" is a marketing technique that intentionally promotes products and services without consumers realizing it.

[1550] A "false review" is a review that contains information that is not true.

[1551] A "summary" is a short summary of the main content of the original text.

[1552] A "personalized summary" is a summary that is optimized based on an individual user's interests and past behavior.

[1553] "Facial recognition technology" is a technology that identifies and analyzes people's faces from images and videos.

[1554] "Voice analysis technology" is a technology that analyzes voice data and analyzes content, emotions, etc.

[1555] An "emotion engine" is a technology that analyzes the user's emotions and optimizes the system's operation based on the results.

[1556] This invention provides a system that enables users to quickly and easily select product reviews on an e-commerce site, and also has the function of recognizing the user's emotions and customizing the review display. Specific embodiments of this system are described below.

[1557] Generating a Program

[1558] This system is composed of a program that combines multiple methods. First, the server collects product reviews based on search queries received from users. To do this, the server retrieves relevant reviews from major e-commerce sites via API and stores them in an internal database.

[1559] Processing Overview

[1560] The following hardware and software are used to implement this program:

[1561] Hardware: Smartphone

[1562] Software: NLP analysis tools (e.g., Google Cloud Natural Language API), emotion analysis tools (e.g., Microsoft Azure Emotion API)

[1563] Database: NoSQL database (e.g. MongoDB)

[1564] Server: AWS Lambda (serverless computing)

[1565] 1. Collect reviews

[1566] The server collects reviews based on the search query submitted by the user. For example, when a user searches for "smartphone case," the server collects related reviews via the e-commerce site's API and stores them in a NoSQL database.

[1567] 2. NLP analysis

[1568] The server uses the Google Cloud Natural Language API to analyze the collected reviews, including classifying reviews as positive, negative, or neutral, extracting key phrases, and analyzing emotional tone, allowing it to quickly provide users with the information they need most.

[1569] 3. Trustworthiness Assessment and Filtering

[1570] The server evaluates the authenticity of the reviews and filters out stealth marketing and fake reviews. This step removes unusual patterns (such as a large number of posts from the same IP address) and reviews with low scores.

[1571] 4. Summary Generation

[1572] The server generates summaries from reliable reviews, which are short summaries of key features of the reviews to help users make a quick decision. For example, a summary might read, "This smartphone case is durable and has a good design, but is a bit pricey."

[1573] 5. Emotion recognition

[1574] The server uses the Microsoft Azure Emotion API to recognize the user's current emotional state. To do this, it uses facial expression analysis using the user's smartphone camera and emotion analysis technology from voice. This allows it to identify the user's emotional state.

[1575] 6. Personalized Display

[1576] The server customizes the display of reviews based on information obtained from the emotion engine. For example, when a user is in a positive emotional state, it highlights positive reviews. This function provides information optimized for each user.

[1577] 7. Display

[1578] Finally, the generated summary is displayed on the user's device. If the device is a smartphone, the appropriate summary is visually presented on the screen.

[1579] Specific examples

[1580] When a user searches for "smartphone case," the server sends the following prompt to the generative AI model:

[1581] Get product-related reviews and categorize them as positive, negative, or neutral based on sentiment. Extract key phrases and generate summaries. Eliminate stealth marketing patterns and assess trustworthiness. Display customized summaries based on the user's emotional state.

[1582] This allows users to quickly obtain reliable review information and enjoy optimized review displays based on their own emotional state and past purchasing history.

[1583] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1584] Step 1:

[1585] A user enters keywords related to a specific product name or category into the search bar and clicks the search button. This input generates a search query to retrieve reviews related to the product. The server receives this search query and secures the data to proceed to the next step.

[1586] Step 2:

[1587] The server collects relevant product reviews from major e-commerce sites based on the received search query. Specifically, it retrieves the relevant reviews via API and stores them in a NoSQL database. For example, it collects hundreds of reviews about "smartphone cases." This allows review information based on the product the user is looking for to be stored in the database.

[1588] Step 3:

[1589] The server applies natural language processing (NLP) technology to analyze the collected product reviews. First, it classifies the review content as positive, negative, or neutral. Next, it extracts important information and evaluation points (key phrases) from each review. It also performs sentiment analysis of the reviews to determine the emotional tone of the review. For example, it distinguishes between reviews that are "very satisfied" and reviews that are "slightly dissatisfied." This allows the information users need to be provided quickly.

[1590] Step 4:

[1591] The server evaluates the credibility of reviews based on the analysis results and filters out stealth marketing and fake reviews. First, it detects unnatural patterns and removes reviews posted in large numbers from the same IP address or reviews with abnormally high ratings. Next, it assigns a credibility score to each review and eliminates reviews with low scores. This step allows reliable reviews to be selected.

[1592] Step 5:

[1593] The server generates easy-to-read summaries from reviews that pass the reliability assessment. Specifically, it summarizes important evaluation points in a concise manner, allowing users to quickly obtain key information without having to read lengthy reviews. For example, a summary might be generated such as, "This smartphone case is highly durable and has a well-received design, but is a little pricey."

[1594] Step 6:

[1595] The server generates personalized summaries based on the user's browsing and purchasing history. It analyzes the types of products the user has previously purchased and provides summaries tailored to those preferences. For example, a user who has previously preferred durable products might receive a summary such as "This case is particularly durable."

[1596] Step 7:

[1597] The server uses an emotion engine to recognize the user's emotions. To do this, it analyzes the user's facial expressions using the device (smartphone) camera and identifies the user's emotional state using the Microsoft Azure Emotion API. It also analyzes the user's vocalizations to recognize the user's emotional state.

[1598] Step 8:

[1599] The server customizes the review display based on the information obtained from the emotion engine and the personalized summary. For example, if the user is in a positive state, it highlights positive reviews, and if the user is in an anxious state, it highlights reliable reviews. This step provides the user with the most appropriate information according to their emotional state.

[1600] Step 9:

[1601] The generated summary is displayed on the device (smartphone) screen. Users can quickly obtain the appropriate summary information to quickly evaluate the product. For example, a summary such as "Many users rate this case highly for its durability and design, and although the price is somewhat high, satisfaction is high" may be displayed.

[1602] 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.

[1603] 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.

[1604] 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.

[1605] 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.

[1606] 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.

[1607] 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.

[1608] 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).

[1609] 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.

[1610] 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."

[1611] 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.

[1612] 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).

[1613] 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.

[1614] 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.

[1615] 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.

[1616] 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.

[1617] 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.

[1618] 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.

[1619] 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.

[1620] 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.

[1621] 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 the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1622] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1623] The following is further disclosed regarding the above embodiment.

[1624] (Claim 1)

[1625] A means for collecting product reviews based on search queries received from users;

[1626] A means for analyzing collected product reviews using natural language processing technology, classifying the reviews, and extracting key phrases;

[1627] A means to assess the authenticity of reviews and filter out stealth marketing and fake reviews;

[1628] a means for generating readable summaries from the analyzed reviews;

[1629] A means for generating personalized reviews based on a user's browsing history and purchasing history;

[1630] means for displaying the generated summary on a user's terminal;

[1631] A system including:

[1632] (Claim 2)

[1633] 2. The system of claim 1, wherein the review classification means classifies reviews into at least three categories: positive, negative, and neutral.

[1634] (Claim 3)

[1635] 10. The system of claim 1, further comprising means for scoring the authoritativeness of a review in response to a user's search query.

[1636] "Example 1"

[1637] (Claim 1)

[1638] A means for collecting product reviews based on search queries received from users;

[1639] A means for analyzing collected product reviews using natural language processing technology, classifying the reviews, and extracting key phrases;

[1640] a means of assessing the authenticity of reviews and filtering fraudulent reviews;

[1641] a means for generating readable summaries from the analyzed reviews;

[1642] means for generating a personalized summary based on a user's browsing and purchasing history;

[1643] means for providing the generated summary on a display device;

[1644] An information processing system including:

[1645] (Claim 2)

[1646] 2. The information processing system according to claim 1, wherein the review classification means classifies reviews into at least three categories: positive, negative, and neutral.

[1647] (Claim 3)

[1648] 10. The information processing system of claim 1, further comprising means for scoring the credibility of reviews in response to a user's search query.

[1649] "Application Example 1"

[1650] (Claim 1)

[1651] A means for collecting product reviews based on search queries received from users;

[1652] A means for analyzing collected product reviews using natural language processing technology, classifying the reviews, and extracting key phrases;

[1653] A means to assess the authenticity of reviews and filter out stealth marketing and fake reviews;

[1654] a means for generating readable summaries from the analyzed reviews;

[1655] A means for generating personalized reviews based on a user's browsing history and purchasing history;

[1656] means for displaying the generated summary on a user's terminal;

[1657] A means of analyzing the content of reviews and classifying them as positive, negative, or neutral;

[1658] A means for analyzing detailed information about the product viewed by the user and generating a prompt sentence;

[1659] A system including:

[1660] (Claim 2)

[1661] 2. The system of claim 1, wherein the review classification means classifies reviews into at least three categories: positive, negative, and neutral.

[1662] (Claim 3)

[1663] 10. The system of claim 1, further comprising means for scoring the authoritativeness of a review in response to a user's search query.

[1664] "Example 2: Combining Emotion Engines"

[1665] (Claim 1)

[1666] A means for collecting product reviews based on search queries received from users;

[1667] A means for analyzing collected product reviews using natural language processing technology, classifying the reviews, and extracting key phrases;

[1668] a means to assess the authenticity of reviews and filter fraudulent reviews;

[1669] a means for generating readable summaries from the analyzed reviews;

[1670] A means for generating personalized reviews based on a user's browsing history and purchasing history;

[1671] means including an emotion engine for recognizing an emotion of a user;

[1672] a means for customizing review display based on emotion recognition results;

[1673] means for displaying the generated summary on a user's terminal;

[1674] A system including:

[1675] (Claim 2)

[1676] 2. The system of claim 1, wherein the review classification means classifies reviews into at least three categories: positive, negative, and neutral.

[1677] (Claim 3)

[1678] 10. The system of claim 1, further comprising means for scoring the authoritativeness of a review in response to a user's search query.

[1679] "Application example 2 when combining emotion engines"

[1680] (Claim 1)

[1681] a means for collecting reviews based on search queries received from users;

[1682] A means for analyzing the collected reviews using natural language processing technology, classifying the reviews, and extracting key phrases;

[1683] A means to assess the authenticity of reviews and filter out stealth marketing and fake reviews;

[1684] a means for generating readable summaries from the analyzed reviews;

[1685] means for generating personalized summaries based on a user's browsing and purchasing history;

[1686] means for analyzing a user's facial expression using facial recognition technology to identify an emotional state;

[1687] A means for analyzing a user's vocalizations using a voice analysis technique to identify an emotional state of the user;

[1688] A way to customize review display based on information obtained from the sentiment engine, and

[1689] means for displaying the generated summary on a user's terminal;

[1690] A system including:

[1691] (Claim 2)

[1692] 2. The system of claim 1, wherein the review classification means classifies reviews into at least three categories: positive, negative, and neutral.

[1693] (Claim 3)

[1694] 10. The system of claim 1, further comprising means for scoring the authoritativeness of a review in response to a user's search query. [Explanation of symbols]

[1695] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means for collecting product reviews based on search queries received from users; A means for analyzing collected product reviews using natural language processing technology, classifying the reviews, and extracting key phrases; A means to assess the authenticity of reviews and filter out stealth marketing and fake reviews; a means for generating readable summaries from the analyzed reviews; A means for generating personalized reviews based on a user's browsing history and purchasing history; means for displaying the generated summary on a user's terminal; A system including:

2. 2. The system according to claim 1, wherein the review classification means classifies reviews into at least three categories: positive, negative, and neutral.

3. The system of claim 1 , further comprising means for scoring the credibility of a review in response to a user's search query.

Images