System

Abstract

A system is provided.SOLUTION: The system includes a means for automatically collecting review information on the Internet, a means for analyzing the collected review information and extracting key keywords and evaluation scores, and a means for generating a review summary on the basis of the analyzed information and displaying the summary through a user interface.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] While there is a large amount of review information on the Internet, this information is fragmented, making it difficult to easily obtain reliable information. Furthermore, when consumers make purchasing decisions, the task of directly reading and interpreting each review and extracting important information takes a great deal of time and effort. As a result, it is difficult to quickly and accurately grasp the evaluation of a product or service. Furthermore, companies lack effective means of accurately grasping the evaluation of their own products and services. Thus, there is a need for a method to efficiently collect and analyze fragmented review information on the Internet and provide reliable summaries. [Means for solving the problem]

[0005] To address these challenges, the present invention provides a system that includes a means for automatically collecting online review information, a means for analyzing the collected review information and extracting key keywords and evaluation scores, and a means for generating review summaries based on the analyzed information and displaying them through a user interface. Specifically, the system efficiently collects review information using a crawler that patrols specific websites, and analyzes the content of the reviews using artificial intelligence and natural language processing technology. This analysis extracts important evaluation points related to products and services and classifies reviews into high- and low-rated reviews. The system then generates and displays summaries in a format that is easily understandable to users, thereby providing fast and reliable information. This system significantly reduces the effort required for consumers to gather information when making purchasing decisions and enables companies to accurately understand the evaluations of their products.

[0006] The "Internet" is a huge global network that interconnects computers and networks around the world.

[0007] "Review information" refers to text and evaluation scores that contain users' evaluations and opinions of products and services.

[0008] "Automatic" means that a computer or machine performs a process on its own without human intervention.

[0009] "Means of collection" is a general term for methods, devices, programs, etc. for collecting information.

[0010] "Means of analysis" is a general term for methods, devices, programs, etc. used to analyze collected data and understand and interpret its contents.

[0011] A "point" is a part of specific data or information that has a particularly important meaning.

[0012] "Keywords" are words that are frequently used in text or data to emphasize a particular meaning.

[0013] "Evaluation score" refers to the score or rating given by a user to a product or service in review information.

[0014] A "summary" is a short summary of a longer piece of text or data.

[0015] A "user interface" is the system or screen through which a user interacts with a computer system or application.

[0016] "Display means" is a general term for methods, devices, programs, etc. for visually presenting data or information on a user interface.

[0017] A "system" is a collection of multiple devices and programs that work together to achieve a specific purpose.

[0018] A "crawler" is a program that automatically crawls websites on the Internet and collects information.

[0019] "Artificial intelligence" refers to computer systems or programs that mimic human intelligence to solve problems and make decisions.

[0020] "Natural language processing technology" refers to the technology that uses computers to analyze, understand, and generate human language (natural language). [Brief explanation of the drawings]

[0021] [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

[0022] 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.

[0023] First, the terms used in the following description will be explained.

[0024] 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).

[0025] 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.

[0026] 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.

[0027] 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.

[0028] 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."

[0029] [First embodiment]

[0030] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0031] 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.

[0032] 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).

[0033] 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.

[0034] 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.

[0035] 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.

[0036] 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.

[0037] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0038] 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.

[0039] 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.

[0040] 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.

[0041] 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."

[0042] System configuration

[0043] The present invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. The configuration and processing of the system are described in detail below.

[0044] Detailed description of the program's processing

[0045] Collecting review information

[0046] The server receives a request from a user to collect review information about a particular product or service.

[0047] Based on the request, the server launches a web crawler, which then visits the specified website (e.g., a major e-commerce site, a review site, etc.) and automatically collects review information.

[0048] The server's crawler extracts the review text, rating score, reviewer information, and posting date and time from each web page, and stores this information in a database.

[0049] Review information analysis

[0050] The server applies natural language processing (NLP) algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, and performing dependency analysis to extract important keywords and phrases.

[0051] The server categorizes each review based on the extracted keywords and phrases, and scores them based on the evaluation score, thereby classifying reviews into positive and negative.

[0052] Generate and display summaries

[0053] The server generates a summary of the reviews based on the analysis, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends.

[0054] In response to a user request, the terminal displays the summary information received from the server on a user interface. The user can refer to this summary to check product and service ratings and make a purchasing decision.

[0055] Specific examples

[0056] 1. Collecting review information

[0057] A user sends a request to a server to collect review information about a particular smartphone.

[0058] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[0059] 2. Review information analysis

[0060] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[0061] The server scores each review based on its rating score and categorizes the reviews as positive and negative, for example identifying 60 out of 80 positive reviews about "battery life."

[0062] 3. Generating and displaying summaries

[0063] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and provides it to the user, including the distribution of rating scores, representative positive and negative opinions, and overall rating trends.

[0064] The device displays this summary information on the user interface, and the user can refer to it to check product reviews. For example, by looking at the summary information, the user can come to a conclusion such as "This smartphone is rated for its long battery life, but there are 20 negative reviews about its camera performance."

[0065] As described above, the system of the present invention can efficiently collect and analyze fragmented review information on the Internet and provide highly reliable information to support user decision-making. This system allows consumers to quickly and accurately understand product and service ratings, and provides companies with an effective means of accurately understanding the ratings of their products and services.

[0066] The processing flow will be explained below.

[0067] Step 1: Submit a request to collect reviews

[0068] A user transmits a request for collecting review information about a specific product or service from a terminal, and the request includes identification information of the product or service (e.g., product ID or name).

[0069] Step 2: Launch the web crawler

[0070] The server receives a request from a user and launches a web crawler, which visits specific pre-defined websites and review sites (e.g., major e-commerce sites, word-of-mouth sites, etc.).

[0071] Step 3: Extracting review information

[0072] The server's crawler analyzes the specified web page and extracts review information, which means analyzing HTML tags and obtaining necessary data such as the review text, rating score, reviewer information, and posting date and time.

[0073] Step 4: Save your review information

[0074] The server stores the extracted review information in a database, adding the website information and product ID from which it was collected as metadata, making the information traceable later.

[0075] Step 5: Begin data analysis

[0076] The server initiates an analysis process of the collected review information based on periodic batch processing or user requests, a procedure for efficiently processing large amounts of data.

[0077] Step 6: Applying Natural Language Processing (NLP)

[0078] The server applies NLP algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, analyzing dependencies, and extracting keywords, thereby identifying important elements of the review.

[0079] Step 7: Categorizing and scoring reviews

[0080] The server categorizes each review based on keywords and phrases extracted through NLP processing, then distinguishes between positive and negative reviews and assigns them a score based on the review's rating.

[0081] Step 8: Generate a summary

[0082] The server generates a summary of the review information based on the analysis results, including the distribution of rating scores, representative positive and negative opinions, and overall rating trends.

[0083] Step 9: Send summary information

[0084] The server sends the generated summary information to the user's terminal for provision to the user. This communication is performed in real time in response to a request or periodically as a batch process.

[0085] Step 10: View summary information

[0086] The terminal displays the received summary information on a user interface, allowing the user to quickly understand the evaluation of a particular product or service and make a purchasing decision.

[0087] The above is the specific program processing flow of this system, which enables users to efficiently obtain highly reliable review information and make decisions.

[0088] Example 1

[0089] 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."

[0090] Until now, much of the review information on the Internet has been scattered individually, making it difficult for users to efficiently compare and analyze it. Additionally, grasping the reliability and overall trends of collected review information has been a time-consuming and labor-intensive process. Furthermore, the lack of a means to appropriately utilize natural language processing technology when analyzing review information has made it difficult to create accurate review summaries. Therefore, there is a need for a system that can efficiently collect and analyze review information and provide highly reliable review summaries.

[0091] 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.

[0092] In this invention, the server includes: means for receiving a request from a user to collect review information about a specific product or service; means for automatically collecting review information from the Internet; means for analyzing the collected review information, tokenizing the review text using natural language processing technology, performing part-of-speech tagging and dependency analysis, and extracting key keywords and evaluation scores; means for generating a summary of the review based on the analyzed information and displaying the summary through a user interface, including the distribution of evaluation scores, representative positive and negative opinions, and overall evaluation trends; and means for displaying the summary information on the user interface in response to a user request. This allows fragmented review information to be efficiently collected and analyzed, enabling users to easily check reviews and make quick and accurate purchasing decisions for products and services.

[0093] The "means for receiving a request from a user to collect review information about a specific product or service" refers to a function in which a user sends a request to collect reviews about a specific product or service to a server and the request is received.

[0094] "Means for automatically collecting review information on the Internet" refers to a function for automatically obtaining review information from specific websites on the Internet, including the use of a crawler program for this purpose.

[0095] "Natural language processing technology" is a technology for analyzing review text, and performs processes such as tokenization, part-of-speech tagging, and dependency analysis.

[0096] "Tokenizing the review text" refers to the process of dividing the entire review text into words and phrases.

[0097] "Part-of-speech tagging" refers to the process of assigning each token the appropriate part of speech (noun, verb, adjective, etc.).

[0098] "Dependency parsing" refers to a parsing technique that reveals the relationships between words in a sentence (e.g., subject and verb, modifier and noun, etc.).

[0099] "Means for extracting key keywords and evaluation scores" refers to the function of extracting important words and phrases from the review text, as well as user evaluations.

[0100] "Generating review summaries" refers to the process of creating information summarizing overall ratings, positive and negative opinions, etc., based on collected and analyzed review information.

[0101] "Means for displaying through a user interface" refers to a function for visually displaying the generated review summary on the user's screen.

[0102] "Rating score distribution" refers to data showing how the rating scores of collected reviews are distributed.

[0103] "Representative positive and negative opinions" refers to the most common positive and negative comments found across all reviews.

[0104] "Overall rating trends" refers to data showing changes in review content and rating scores over a certain period of time.

[0105] A "crawler" is a program that automatically crawls specific websites on the Internet and collects specified information.

[0106] The present invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. The main feature of this system is that it automatically collects review information based on user requests, analyzes it using natural language processing technology, and displays it as a summary. Specific embodiments are described below.

[0107] System configuration

[0108] This system is composed of three main components: a server, a terminal, and a user.

[0109] 1. Server

[0110] The server receives a request from a user to collect review information about a specific product or service. Upon receiving the request, the server launches a web crawler program using the Python Scrapy library to collect review information from the specified website. The collected review information (e.g., review text, rating score, reviewer information, and posting date and time) is stored in a database such as MySQL or MongoDB.

[0111] 2. Review information analysis

[0112] The server applies natural language processing (NLP) techniques to the collected review information. Using Python's NLTK and spaCy libraries, it tokenizes the review text, tags it with parts of speech, and performs dependency analysis to extract important keywords and phrases. For example, keywords such as "battery life" and "camera performance" are extracted.

[0113] 3. Review categorization and scoring

[0114] The server categorizes the reviews based on the analysis results and scores them as positive or negative based on their rating. This information is used to generate a comprehensive review summary, which includes the distribution of rating scores, representative examples of positive and negative comments, and the overall rating trend.

[0115] 4. Terminals and User Interfaces

[0116] In response to a user request, the device displays the review summary information received from the server on the user interface. The user refers to this summary to check the overall evaluation of a specific product or service and make a purchasing decision. As a specific example, the device screen may display the following message: "This smartphone has many positive reviews regarding its battery life, but there are 20 negative comments regarding its camera performance."

[0117] Prompt Sentence Examples

[0118] Below is a specific example of a request that a user sends to the server.

[0119] "Collect and analyze reviews of the latest smartphones and generate summaries."

[0120] "Please categorize the positive and negative reviews of a particular cosmetic product and give us a representative opinion for each."

[0121] The system of the present invention allows users to efficiently collect fragmented review information on the Internet and obtain highly reliable information based on the analysis results. This allows users to quickly and accurately understand product and service evaluations and make wise purchasing decisions. It also provides companies with an effective means of accurately understanding the evaluations of their own products and services.

[0122] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0123] Step 1:

[0124] A user sends a request to the server to collect review information about a particular product or service. The input is the product or service information specified by the user, including a prompt such as "Please collect reviews about the latest smartphones." The output is a response confirming that the server received the request.

[0125] Step 2:

[0126] The server launches a web crawler using the Python Scrapy library based on the request received from the user. The input is the product or service information in the user request and a list of websites to be collected. The output is log information indicating that the crawler has started accessing the websites to be collected.

[0127] Step 3:

[0128] The server's web crawler visits specified websites (e.g., major e-commerce sites or review sites) and collects review information (review text, rating score, reviewer information, and posting date and time). The input is the HTML data of the web pages accessed by the crawler. The output is a dataset of extracted review information, which is stored in a database such as MySQL or MongoDB.

[0129] Step 4:

[0130] The server applies natural language processing (NLP) to the database containing the collected review information. It uses Python's NLTK and spaCy libraries to tokenize the review text and perform part-of-speech tagging and dependency analysis. The input is the review text data in the database. The output is a dataset containing the analyzed tokens, part-of-speech tags, and dependencies.

[0131] Step 5:

[0132] The server extracts important keywords and phrases based on the analysis results and categorizes each review into a category. It also scores reviews as positive or negative based on their rating scores. The input is the dataset after NLP analysis. The output is a dataset of reviews categorized and scored.

[0133] Step 6:

[0134] The server generates a review summary based on the scored review information. This summary includes the distribution of rating scores, representative examples of positive and negative opinions, and overall rating trends. The input is the categorized and scored review information. The output is the generated review summary.

[0135] Step 7:

[0136] The device displays review summary information received from the server on the user interface in response to a user request. The input is the review summary sent from the server. The output is the review summary displayed on the user interface, which the user can refer to to check the evaluation of the product or service. Specifically, the device displays the message, "This smartphone has many positive reviews regarding its battery life, but there are 20 negative comments regarding its camera performance."

[0137] (Application example 1)

[0138] 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."

[0139] Conventional online review information collection and analysis systems have made it difficult for users to accurately and efficiently collect and understand review information about specific products. This requires manually examining a huge number of reviews, which is time-consuming and labor-intensive. Furthermore, there are limited means of providing reliable review information in a visually easy-to-understand format. Therefore, there is a need for technology that can more quickly and accurately assist users in making purchasing decisions.

[0140] 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.

[0141] In this invention, the server includes means for automatically collecting evaluation information on the Internet, means for analyzing the collected evaluation information and extracting key keywords and evaluation scores, means for generating an evaluation summary based on the analyzed information and displaying it through a user interface, input means for users to search for evaluation information on a specific product, and means for generating a summary based on the evaluation information and displaying it in a visually easy-to-understand format, thereby enabling users to quickly and efficiently obtain and understand highly reliable evaluation information on a specific product.

[0142] The "Internet" is a global network that interconnects computers and networks around the world, enabling the exchange of information.

[0143] "Evaluation information" refers to data such as users' opinions and impressions about products and services, and evaluation scores.

[0144] "Means" refers to methods, techniques, devices, etc. used to achieve a certain purpose.

[0145] "Collection" is the act of gathering information or data.

[0146] "Analysis" is the process of examining collected information and data in detail to clarify its structure, meaning, and relationships.

[0147] "Points" refers to particularly important elements or main points.

[0148] A "keyword" refers to an important word that characterizes a particular piece of information.

[0149] "Evaluation score" refers to data that quantifies the evaluation of a product or service.

[0150] "Summary" refers to a concise report or overview that summarises detailed information.

[0151] "User interface" refers to the means or environment through which a user and a system interact.

[0152] "Input" is the act of providing data or instructions to a system.

[0153] "Visual" refers to a form or expression that can be perceived by the eye.

[0154] A "website" is a collection of information hosted on the Internet.

[0155] A "web crawler" is a program that automatically crawls web pages on the Internet and collects information.

[0156] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[0157] "Positive" refers to positive opinions and evaluations.

[0158] "Negative" refers to negative opinions or evaluations.

[0159] A "trend" refers to a movement or change that shows a particular direction or tendency.

[0160] This invention is a system that automatically collects and analyzes evaluation information related to specific products and services from the Internet, and generates easy-to-understand summaries based on the results to support users' purchasing decisions. This system is mainly composed of a server and user terminals.

[0161] 1. System Configuration

[0162] Hardware

[0163] 1. Server

[0164] Cloud servers (e.g., AWS, GCP) are used to collect, analyze, and generate summaries of evaluation information.

[0165] 2. User Device

[0166] It is designed for smartphones and tablets and is responsible for displaying summary information.

[0167] software

[0168] 1. Web crawler

[0169] Using tools such as Scrapy, evaluation information is automatically collected from specific websites on the Internet.

[0170] 2. Natural Language Processing (NLP)

[0171] Using tools such as NLTK or Spacy, the collected evaluation information is analyzed to extract keywords and evaluation scores.

[0172] 3. Database

[0173] Use MySQL or MongoDB to store the collected and analyzed data.

[0174] 4. Front-end

[0175] It uses React Native to display summary information in a user-friendly interface.

[0176] 5. Backend

[0177] Use Django to coordinate data processing and user interface processing.

[0178] 2. System operation explanation

[0179] Collection Phase

[0180] A user submits a search request for rating information about a particular product.

[0181] The server receives the request and launches a web crawler, which then visits the specified website, collects rating information (e.g., review text, rating score, reviewer information, and posting date and time), and stores it in a database.

[0182] Analysis Phase

[0183] The server analyzes the rating information stored in the database using NLP techniques, including tokenization, part-of-speech tagging, and dependency analysis to extract important keywords and phrases.

[0184] Based on the extracted keywords and phrases, the rating information is categorized and positive and negative reviews are distinguished.

[0185] Summary Generation Phase

[0186] The server generates a summary of the review information based on the analysis results, including the distribution of review scores, representative examples of positive and negative comments, and overall review trends.

[0187] The user terminal displays the summary information through a user interface, allowing the user to quickly check the evaluation of the product or service.

[0188] 3. Examples and prompts

[0189] Examples:

[0190] Let's say a user wants to find out reviews about "XYZ Smartphone."

[0191] Users search for "XYZ smartphone" in the app, and the server collects, analyzes, and generates summaries of related reviews from various websites.

[0192] The analysis results in a conclusion such as, "This smartphone is rated for its long battery life, but there are 20 negative reviews about its camera performance."

[0193] Example prompt sentence:

[0194] Please analyze online reviews for XYZ smartphone and tell me the number of positive and negative reviews and what the representative opinions are.

[0195] The present invention allows users to quickly and efficiently obtain highly reliable evaluation information, which can support their own purchasing decisions.

[0196] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0197] Step 1:

[0198] User enters a search request

[0199] When a user wants to know evaluation information about a specific product, the user starts up an app on a device such as a smartphone or tablet and inputs the name of the product.

[0200] Input: Specific product name

[0201] Output: Search request

[0202] Step 2:

[0203] The device sends a request to the server

[0204] After the terminal accepts the user input, it sends the search request to the server.

[0205] Input: Search request

[0206] Output: Request sent to server

[0207] Step 3:

[0208] The server launches the web crawler

[0209] After the server receives the search request, it launches a web crawler to collect reputation information from the specified website.

[0210] Input: Search request

[0211] Output: Collected rating information (review text, rating score, reviewer information, posting date and time)

[0212] Step 4:

[0213] The server stores the rating information in a database

[0214] The server stores the collected evaluation information in a database.

[0215] Input: Collected evaluation information

[0216] Output: Evaluation information stored in a database

[0217] Step 5:

[0218] The server analyzes the evaluation information.

[0219] The server retrieves the rating information from the database and analyzes it using natural language processing techniques, including tokenization, part-of-speech tagging, and dependency analysis.

[0220] Input: Rating information retrieved from the database

[0221] Output: Analyzed data (keywords, rating scores, reviewer information, etc.)

[0222] Step 6:

[0223] The server categorizes positive and negative reviews

[0224] Based on the analyzed data, the server categorizes the rating information and distinguishes between positive and negative reviews.

[0225] Input: Parsed data

[0226] Output: Categorized positive and negative reviews

[0227] Step 7:

[0228] The server generates a summary of the rating information

[0229] The server generates a summary of the rating information based on representative examples of positive and negative reviews, the distribution of rating scores, and overall rating trends.

[0230] Input: Categorized positive and negative reviews

[0231] Output: Summary of evaluation information

[0232] Step 8:

[0233] User terminal displays summary information

[0234] When the user accesses the app again, it receives a summary of the rating information from the server and displays it visually through the user interface.

[0235] Input: Summary of rating information sent from the server

[0236] Output: A summary of the evaluation information displayed on the terminal.

[0237] 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.

[0238] System configuration

[0239] This invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. By combining this system with an emotion engine, it has the ability to recognize user emotions from review information and generate emotion scores. The following describes the system's configuration and processing in detail.

[0240] Detailed description of the program's processing

[0241] Collecting review information

[0242] The server receives a request from a user to collect review information about a particular product or service.

[0243] Based on the request, the server launches a web crawler, which then visits the specified website (e.g., a major e-commerce site, a review site, etc.) and automatically collects review information.

[0244] The server's crawler extracts the review text, rating score, reviewer information, and posting date and time from each web page, and stores this information in a database.

[0245] Review information analysis

[0246] The server applies natural language processing (NLP) algorithms to the collected review information, including tokenizing the review text, tagging parts of speech, dependency analysis, and keyword extraction, to identify important elements of the review.

[0247] The server categorizes each review based on the extracted keywords and phrases, and scores them based on the evaluation score, thereby classifying reviews into positive and negative.

[0248] Emotion recognition by emotion engine

[0249] The server uses an emotion engine to recognize the user's emotion from the analyzed review information. The emotion engine determines the emotion (e.g., joy, anger, sadness, etc.) of the user who wrote the review based on the context and keywords in the review text.

[0250] The server quantifies the emotions determined by the emotion engine and generates an emotion score, which is reflected in the review summary along with the positive / negative rating.

[0251] Generate and display summaries

[0252] The server generates a summary of the review information based on the analysis results and sentiment scores, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends, as well as the sentiment scores.

[0253] In response to a user request, the terminal displays the summary information received from the server on a user interface. The user can refer to this summary to check product and service ratings and make a purchasing decision.

[0254] Specific examples

[0255] 1. Collecting review information

[0256] A user sends a request to a server to collect review information about a particular smartphone.

[0257] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[0258] 2. Review information analysis

[0259] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[0260] The server scores each review based on its rating score and categorizes the reviews as positive and negative, for example identifying 60 out of 80 positive reviews about "battery life."

[0261] 3. Emotion Recognition by Emotion Engine

[0262] The server uses an emotion engine to analyze the review text and recognize the user's emotion. For example, it recognizes the emotion "joy" from a review text such as "very satisfied" and generates an emotion score.

[0263] The server stores the sentiment scores along with other analytical results in a database, recording the emotional state of each review.

[0264] 4. Generating and displaying summaries

[0265] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and sentiment scores and provides it to the user. This summary includes the distribution of rating scores, representative positive and negative opinions, overall rating trends, and sentiment scores.

[0266] The device displays this summary information on the user interface, allowing users to check product ratings. For example, users can learn that "this smartphone has a good battery life, but the camera performance has 20 negative reviews," and also get a conclusion that "user sentiment scores are high overall."

[0267] As described above, the system of the present invention efficiently collects and analyzes fragmented review information on the Internet and recognizes user sentiment to provide highly reliable information. This system allows users to quickly and accurately understand product and service ratings, and provides companies with an effective means of accurately understanding the ratings of their products and services.

[0268] The processing flow will be explained below.

[0269] Step 1: Submit a request to collect reviews

[0270] A user transmits a request for collecting review information about a specific product or service from a terminal, and the request includes identification information of the product or service (e.g., product ID or name).

[0271] Step 2: Launch the web crawler

[0272] The server receives a request from a user and launches a web crawler, which visits specific pre-defined websites and review sites (e.g., major e-commerce sites, word-of-mouth sites, etc.).

[0273] Step 3: Extracting review information

[0274] The server's crawler analyzes the specified web page and extracts review information, which means analyzing HTML tags and obtaining necessary data such as the review text, rating score, reviewer information, and posting date and time.

[0275] Step 4: Save your review information

[0276] The server stores the extracted review information in a database, adding the website information and product ID from which it was collected as metadata, making the information traceable later.

[0277] Step 5: Begin data analysis

[0278] The server initiates an analysis process of the collected review information based on periodic batch processing or user requests, a procedure for efficiently processing large amounts of data.

[0279] Step 6: Applying Natural Language Processing (NLP)

[0280] The server applies NLP algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, analyzing dependencies, and extracting keywords, thereby identifying important elements of the review.

[0281] Step 7: Categorizing and scoring reviews

[0282] The server categorizes each review based on keywords and phrases extracted through NLP processing, then distinguishes between positive and negative reviews and assigns them a score based on the review's rating.

[0283] Step 8: Emotion Recognition with the Emotion Engine

[0284] The server uses an emotion engine to recognize the user's emotion from the analyzed review information. The emotion engine determines the emotion (e.g., joy, anger, sadness, etc.) of the user who wrote the review based on the context and keywords in the review text.

[0285] The server quantifies the emotions determined by the emotion engine and generates an emotion score, which is reflected in the review summary along with the positive / negative rating.

[0286] Step 9: Generate a summary

[0287] The server generates a summary of the review information based on the analysis results and sentiment scores, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends, as well as the sentiment scores.

[0288] Step 10: Send summary information

[0289] The server sends the generated summary information to the user's terminal for provision to the user. This communication is performed in real time in response to a request or periodically as a batch process.

[0290] Step 11: View summary information

[0291] The terminal displays the received summary information on a user interface, allowing the user to quickly understand the evaluation of a particular product or service and make a purchasing decision.

[0292] As described above, this system efficiently collects and analyzes online review information and provides highly reliable information that also reflects user sentiment, allowing users to quickly and accurately understand product and service evaluations and make appropriate decisions.

[0293] Example 2

[0294] 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."

[0295] Currently, there is a huge amount of review information on the Internet, but this information is fragmented, and checking each review individually is extremely time-consuming. This makes it difficult for users to quickly and accurately grasp the overall evaluation of a product or service. Furthermore, reviews contain many emotional elements, which can significantly influence the evaluation. However, there is a lack of systems that can properly recognize and reflect emotions. This makes it difficult to provide reliable review summaries.

[0296] 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 a means for automatically collecting review information on the Internet, a means for analyzing the collected review information and extracting key keywords and evaluation scores, a means for recognizing the user's emotions based on the analyzed information and generating an emotion score, and a means for generating a summary of the review and displaying it through a user interface. This allows the user to easily grasp a reliable comprehensive evaluation that includes emotional elements without having to check each piece of fragmented information one by one.

[0297] The "Internet" is an information and communications infrastructure that interconnects computer networks around the world.

[0298] "Review information" is data such as sentences and scores written by users evaluating products or services.

[0299] "Means of collection" refers to the method of automatically obtaining information from the Internet using programs such as web crawlers.

[0300] "Means of analysis" refers to a method of using natural language processing technology to analyze collected data and extract important information.

[0301] "Keywords" refer to important words or phrases extracted from the review text.

[0302] A "rating score" is a numerical rating given by a user to a product or service.

[0303] The "emotion score" is an index that quantifies the emotions of the user who wrote the review by analyzing the review text.

[0304] A "summary" is information that aggregates and summarizes the analysis results of multiple reviews.

[0305] A "user interface" refers to the screen and operating means that allow a user to interact with a system.

[0306] A "crawler" is a program that automatically visits multiple web pages and collects information.

[0307] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[0308] This invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. By combining this system with an emotion engine, it has the ability to recognize user emotions from review information and generate emotion scores. The following describes the system's configuration and processing in detail.

[0309] System configuration

[0310] The system consists of three main components: a server, a terminal, and a user.

[0311] server

[0312] The server plays a key role in collecting and analyzing review information. The server has the following functions:

[0313] Collection function: Launches a web crawler and collects review information from specified e-commerce sites and review sites.

[0314] Analysis function: Using natural language processing (NLP) technology, collected review information is analyzed and keywords and rating scores are extracted.

[0315] Emotion Recognition: An emotion engine is used to recognize user emotions from the review text and generate an emotion score.

[0316] Summary generation function: Generates a summary of review information based on analysis results and sentiment scores.

[0317] Terminal

[0318] The terminal provides an interface that receives requests from users and displays summary information. The terminal has the following functions:

[0319] Request reception function: Receives collection requests from users and sends them to the server.

[0320] Summary display function: Displays the summary information received from the server on the user interface.

[0321] User

[0322] Users can check review information about products and services through their terminals and make purchasing decisions.

[0323] Hardware and software used

[0324] Web crawler: A program that crawls designated websites on the Internet to collect review information.

[0325] Natural language processing (NLP) technology: Technology that analyzes collected review information and extracts important keywords and phrases.

[0326] Sentiment engine: A software tool that recognizes user emotions from review text and generates an emotion score.

[0327] Database: A data storage system for storing the collected and analyzed review information and generated sentiment scores.

[0328] Specific examples

[0329] 1. Collecting review information

[0330] A user sends a request to a server to collect review information about a particular smartphone.

[0331] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[0332] 2. Review information analysis

[0333] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[0334] 3. Emotion Recognition by Emotion Engine

[0335] The server uses an emotion engine to analyze the review text and recognize the user's emotion. For example, it recognizes the emotion "joy" from a review text such as "very satisfied" and generates an emotion score.

[0336] 4. Generating and displaying summaries

[0337] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and sentiment scores and provides it to the user. This summary includes the distribution of rating scores, representative positive and negative opinions, overall rating trends, and sentiment scores.

[0338] The device displays this summary information on the user interface, allowing users to check product ratings. For example, users can learn that "this smartphone has a good battery life, but the camera performance has 20 negative reviews," and also get a conclusion that "user sentiment scores are high overall."

[0339] Through the above processing steps, the system provides users with reliable review information and analysis results.

[0340] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0341] Step 1: Accepting a user request

[0342] A user requests the collection of review information about a specific product or service from a terminal.

[0343] Input: User request (e.g., "Collect review information for smartphone A")

[0344] The terminal receives the user's request and transmits the request to the server.

[0345] Output: Request data to the server

[0346] Step 2: Collect review information

[0347] The server launches a web crawler based on a request received from a user.

[0348] Input: User request data

[0349] The server's web crawler patrols designated e-commerce sites and review sites and automatically collects relevant review information.

[0350] What happens: The server configures the web crawler and performs the task of gathering reviews for the specified sites.

[0351] Output: Collected review information (review text, rating score, reviewer information, posting date and time)

[0352] Step 3: Analyze review information

[0353] The server applies natural language processing (NLP) techniques to the review information stored in the database.

[0354] Input: Saved review information data

[0355] The server tokenizes (divides) the review text into words, tags it with parts of speech, and performs dependency analysis to extract important keywords and phrases.

[0356] How it works: The server uses an NLP library to tokenize and parse specific words in the review text, assigning them parts of speech, and also performs contextual analysis of the review text to extract important phrases.

[0357] Output: Parsed keywords, rating score, reviewer information

[0358] Step 4: Emotion Recognition with the Emotion Engine

[0359] The server recognizes the user's emotions from the analyzed review information using an emotion engine.

[0360] Input: Analyzed keywords, rating score, reviewer information

[0361] The server determines the sentiment based on the context and keywords in the review text and generates a numerical sentiment score.

[0362] Specific operation: The server uses an emotion recognition algorithm to perform contextual analysis of the review text and quantify the type of emotion (e.g., joy, anger) and its intensity.

[0363] Output: Sentiment score

[0364] Step 5: Generate a summary

[0365] The server generates a summary of the review information based on the analysis results and the sentiment score.

[0366] Input: Analysis results, emotion score

[0367] The server generates a summary that includes the distribution of rating scores, representative examples of positive and negative comments, overall rating trends, and sentiment scores.

[0368] What it does: The server combines the analysis results with sentiment scores, aggregates the data, and generates a visualizable summary. It also extracts representative opinions from each review and analyzes overall rating trends.

[0369] Output: Generated review summary

[0370] Step 6: View the summary

[0371] When a user requests summary information, the server sends the summary to the terminal.

[0372] Input: User summary information request

[0373] The server transmits the summary information generated in response to the request to the terminal.

[0374] The terminal displays the received summary information on a user interface.

[0375] Specific operation: The terminal analyzes the summary information received from the server and displays it visually in an easy-to-understand manner on the user interface.

[0376] Output: Review summary displayed to the user (e.g., distribution of individual rating scores, positive / negative opinions, sentiment score)

[0377] Through the above processing steps, the system provides users with reliable review information and analysis results.

[0378] (Application example 2)

[0379] 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."

[0380] Conventional methods for evaluating products based on online reviews have been difficult to utilize effectively due to the sheer volume of review information. Furthermore, simply averaging the evaluation scores can miss important information and users' feelings contained in the review text. In order for users to evaluate products and services and make appropriate purchasing decisions, it is necessary to efficiently analyze review information and provide reliable summary information.

[0381] 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 automatically collecting review information from the Internet, means for analyzing the collected review information and extracting key keywords and evaluation scores, means for generating review summaries based on the analyzed information and displaying them through a user interface, and means including an emotion recognition engine for generating emotion scores from the review text and reflecting the generated emotion scores in the summaries. This allows users to grasp not only the overall evaluation of the review information but also the emotional trends contained in each review, enabling reliable product evaluations based on the review information.

[0382] "Online review information" refers to user evaluations and opinions of products and services posted on the Internet.

[0383] "Automatic collection means" refers to a device or program that automatically collects designated information without requiring manual operation by the user.

[0384] "Analysis means" refers to the functions and algorithms used to decipher the collected data and understand its contents.

[0385] "Key keywords" refer to words or phrases that have particular significance in the review text.

[0386] "Rating score" refers to the numerical rating given by users in reviews of products or services.

[0387] "Means for generating summaries" refers to the function for creating summarized information based on collected and analyzed information.

[0388] "Means of displaying through a user interface" refers to the screens and operating methods used to visually present information to the user.

[0389] An "emotion recognition engine" refers to technology or algorithms that identify and quantify emotions in documents through text analysis.

[0390] An "emotion score" is a numerical value that quantitatively represents the emotion contained in a sentence.

[0391] A "generative AI model" refers to an artificial intelligence model built to learn from data and perform a specified task (in this case, sentiment analysis or review analysis).

[0392] The present invention is a system that automatically collects and analyzes review information available on the Internet and provides users with highly reliable review summaries. This system has the function of generating emotion scores from reviews using an emotion recognition engine and reflecting these emotion scores in summaries. The detailed configuration and operation of the system are described below.

[0393] System configuration

[0394] The system mainly consists of the following components:

[0395] 1. Server:

[0396] How to collect review information: The server launches a web crawler to automatically collect review information from specific websites. This crawls through specified web pages and extracts data such as the review text, rating score, reviewer information, and posting date and time.

[0397] Analysis method: We use natural language processing (NLP) technology to analyze the collected review information, specifically tokenization, part-of-speech tagging, dependency analysis, and keyword extraction.

[0398] Emotion Recognition Engine: Recognizes emotions from the review text and generates an emotion score. The emotion engine determines the user's emotion (e.g., joy, anger, sadness, etc.) based on the context and keywords in the review text.

[0399] 2. Terminal:

[0400] Summary generation method: The server analyzes the data and generates a sentiment score to generate a reliable review summary, which includes the distribution of rating scores, representative examples of positive and negative opinions, overall rating trends, and sentiment scores.

[0401] User interface: The user can view and refer to the review summary received from the server through the terminal.

[0402] Software and hardware used in the implementation

[0403] Hardware: Smartphones, tablets, PCs

[0404] software:

[0405] Python: a programming language

[0406] BeautifulSoup: A library for web crawling

[0407] requests: HTTP request library

[0408] nltk: A natural language processing library

[0409] SentimentIntensityAnalyzer: An nltk module for sentiment analysis.

[0410] Specific examples

[0411] If a user wants to check reviews of the latest smartphones, they first search for "latest smartphones" within the app. This search request is sent to a server, which collects review information from the specified online shopping site. The collected reviews are analyzed using natural language processing technology to extract important keywords and evaluation scores. An emotion recognition engine is also used to generate an emotion score from the review text. Finally, a reliable review summary is generated based on the analysis results and emotion score, and is displayed through the device's user interface.

[0412] Prompt Sentence Examples

[0413] "Collect reviews of this smartphone and display an overall rating summary based on the analysis results and sentiment score."

[0414] In this way, users can understand not only the overall evaluation of the review information but also the emotional trends contained in each review, enabling them to make highly reliable purchasing decisions.

[0415] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0416] Step 1:

[0417] A user sends a request to the server from their device to collect reviews about a specific product. The input is the user's request, including the product name and category. The server receives this request.

[0418] Step 2:

[0419] The server launches a web crawler and crawls designated websites to collect review information. The input is the user's request, and the output is the collected review information. Specifically, the server uses the crawler to analyze web pages and extract the review text, rating score, reviewer information, and posting date and time.

[0420] Step 3:

[0421] The server stores the collected review information in a database. The input is the collected review information, and the output is analyzable data stored in the database. The server stores the review information by dividing it into fields.

[0422] Step 4:

[0423] The server applies natural language processing (NLP) techniques to the stored review information. The input is the review information read from the database, and the output is the parsed review content. Specifically, the server tokenizes each review text and performs part-of-speech tagging, dependency analysis, keyword extraction, etc.

[0424] Step 5:

[0425] The server categorizes reviews based on the results of natural language processing and categorizes positive and negative reviews based on their rating scores. The input is the analyzed review content, and the output is the categorized reviews. The server also calculates the distribution of rating scores.

[0426] Step 6:

[0427] The server uses an emotion recognition engine to analyze the review text and generate an emotion score. The input is the categorized review content, and the output is an emotion score for each review. Specifically, the server calculates an emotion score (e.g., joy, anger, sadness, etc.) based on the context and keywords in the review text.

[0428] Step 7:

[0429] The server generates a review summary based on the analysis results and sentiment scores. The input is the review analysis results and sentiment scores, and the output is a review summary. The server creates a comprehensive summary that includes the distribution of rating scores, representative examples of positive and negative opinions, overall rating trends, and sentiment scores.

[0430] Step 8:

[0431] When a user requests summary information, the server generates a review summary and sends it to the terminal. The input is the user's summary request, and the output is the summary information.

[0432] Step 9:

[0433] The terminal displays the review summaries received from the server on the user interface. The input is the summary information sent from the server, and the output is the review summary displayed on the user interface. This allows the user to refer to the product ratings and make a purchasing decision.

[0434] 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.

[0435] 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.

[0436] 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.

[0437] [Second embodiment]

[0438] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0439] 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.

[0440] 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).

[0441] 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.

[0442] 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.

[0443] 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).

[0444] 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.

[0445] 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.

[0446] 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.

[0447] 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.

[0448] 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.

[0449] 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."

[0450] System configuration

[0451] The present invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. The configuration and processing of the system are described in detail below.

[0452] Detailed description of the program's processing

[0453] Collecting review information

[0454] The server receives a request from a user to collect review information about a particular product or service.

[0455] Based on the request, the server launches a web crawler, which then visits the specified website (e.g., a major e-commerce site, a review site, etc.) and automatically collects review information.

[0456] The server's crawler extracts the review text, rating score, reviewer information, and posting date and time from each web page, and stores this information in a database.

[0457] Review information analysis

[0458] The server applies natural language processing (NLP) algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, and performing dependency analysis to extract important keywords and phrases.

[0459] The server categorizes each review based on the extracted keywords and phrases, and scores them based on the evaluation score, thereby classifying reviews into positive and negative.

[0460] Generate and display summaries

[0461] The server generates a summary of the reviews based on the analysis, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends.

[0462] In response to a user request, the terminal displays the summary information received from the server on a user interface. The user can refer to this summary to check product and service ratings and make a purchasing decision.

[0463] Specific examples

[0464] 1. Collecting review information

[0465] A user sends a request to a server to collect review information about a particular smartphone.

[0466] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[0467] 2. Review information analysis

[0468] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[0469] The server scores each review based on its rating score and categorizes the reviews as positive and negative, for example identifying 60 out of 80 positive reviews about "battery life."

[0470] 3. Generating and displaying summaries

[0471] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and provides it to the user, including the distribution of rating scores, representative positive and negative opinions, and overall rating trends.

[0472] The device displays this summary information on the user interface, and the user can refer to it to check product reviews. For example, by looking at the summary information, the user can come to a conclusion such as "This smartphone is rated for its long battery life, but there are 20 negative reviews about its camera performance."

[0473] As described above, the system of the present invention can efficiently collect and analyze fragmented review information on the Internet and provide highly reliable information to support user decision-making. This system allows consumers to quickly and accurately understand product and service ratings, and provides companies with an effective means of accurately understanding the ratings of their products and services.

[0474] The processing flow will be explained below.

[0475] Step 1: Submit a request to collect reviews

[0476] A user transmits a request for collecting review information about a specific product or service from a terminal, and the request includes identification information of the product or service (e.g., product ID or name).

[0477] Step 2: Launch the web crawler

[0478] The server receives a request from a user and launches a web crawler, which visits specific pre-defined websites and review sites (e.g., major e-commerce sites, word-of-mouth sites, etc.).

[0479] Step 3: Extracting review information

[0480] The server's crawler analyzes the specified web page and extracts review information, which means analyzing HTML tags and obtaining necessary data such as the review text, rating score, reviewer information, and posting date and time.

[0481] Step 4: Save your review information

[0482] The server stores the extracted review information in a database, adding the website information and product ID from which it was collected as metadata, making the information traceable later.

[0483] Step 5: Begin data analysis

[0484] The server initiates an analysis process of the collected review information based on periodic batch processing or user requests, a procedure for efficiently processing large amounts of data.

[0485] Step 6: Applying Natural Language Processing (NLP)

[0486] The server applies NLP algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, analyzing dependencies, and extracting keywords, thereby identifying important elements of the review.

[0487] Step 7: Categorizing and scoring reviews

[0488] The server categorizes each review based on keywords and phrases extracted through NLP processing, then distinguishes between positive and negative reviews and assigns them a score based on the review's rating.

[0489] Step 8: Generate a summary

[0490] The server generates a summary of the review information based on the analysis results, including the distribution of rating scores, representative positive and negative opinions, and overall rating trends.

[0491] Step 9: Send summary information

[0492] The server sends the generated summary information to the user's terminal for provision to the user. This communication is performed in real time in response to a request or periodically as a batch process.

[0493] Step 10: View summary information

[0494] The terminal displays the received summary information on a user interface, allowing the user to quickly understand the evaluation of a particular product or service and make a purchasing decision.

[0495] The above is the specific program processing flow of this system, which enables users to efficiently obtain highly reliable review information and make decisions.

[0496] Example 1

[0497] 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."

[0498] Until now, much of the review information on the Internet has been scattered individually, making it difficult for users to efficiently compare and analyze it. Additionally, grasping the reliability and overall trends of collected review information has been a time-consuming and labor-intensive process. Furthermore, the lack of a means to appropriately utilize natural language processing technology when analyzing review information has made it difficult to create accurate review summaries. Therefore, there is a need for a system that can efficiently collect and analyze review information and provide highly reliable review summaries.

[0499] 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.

[0500] In this invention, the server includes: means for receiving a request from a user to collect review information about a specific product or service; means for automatically collecting review information from the Internet; means for analyzing the collected review information, tokenizing the review text using natural language processing technology, performing part-of-speech tagging and dependency analysis, and extracting key keywords and evaluation scores; means for generating a summary of the review based on the analyzed information and displaying the summary through a user interface, including the distribution of evaluation scores, representative positive and negative opinions, and overall evaluation trends; and means for displaying the summary information on the user interface in response to a user request. This allows fragmented review information to be efficiently collected and analyzed, enabling users to easily check reviews and make quick and accurate purchasing decisions for products and services.

[0501] The "means for receiving a request from a user to collect review information about a specific product or service" refers to a function in which a user sends a request to collect reviews about a specific product or service to a server and the request is received.

[0502] "Means for automatically collecting review information on the Internet" refers to a function for automatically obtaining review information from specific websites on the Internet, including the use of a crawler program for this purpose.

[0503] "Natural language processing technology" is a technology for analyzing review text, and performs processes such as tokenization, part-of-speech tagging, and dependency analysis.

[0504] "Tokenizing the review text" refers to the process of dividing the entire review text into words and phrases.

[0505] "Part-of-speech tagging" refers to the process of assigning each token the appropriate part of speech (noun, verb, adjective, etc.).

[0506] "Dependency parsing" refers to a parsing technique that reveals the relationships between words in a sentence (e.g., subject and verb, modifier and noun, etc.).

[0507] "Means for extracting key keywords and evaluation scores" refers to the function of extracting important words and phrases from the review text, as well as user evaluations.

[0508] "Generating review summaries" refers to the process of creating information summarizing overall ratings, positive and negative opinions, etc., based on collected and analyzed review information.

[0509] "Means for displaying through a user interface" refers to a function for visually displaying the generated review summary on the user's screen.

[0510] "Rating score distribution" refers to data showing how the rating scores of collected reviews are distributed.

[0511] "Representative positive and negative opinions" refers to the most common positive and negative comments found across all reviews.

[0512] "Overall rating trends" refers to data showing changes in review content and rating scores over a certain period of time.

[0513] A "crawler" is a program that automatically crawls specific websites on the Internet and collects specified information.

[0514] The present invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. The main feature of this system is that it automatically collects review information based on user requests, analyzes it using natural language processing technology, and displays it as a summary. Specific embodiments are described below.

[0515] System configuration

[0516] This system is composed of three main components: a server, a terminal, and a user.

[0517] 1. Server

[0518] The server receives a request from a user to collect review information about a specific product or service. Upon receiving the request, the server launches a web crawler program using the Python Scrapy library to collect review information from the specified website. The collected review information (e.g., review text, rating score, reviewer information, and posting date and time) is stored in a database such as MySQL or MongoDB.

[0519] 2. Review information analysis

[0520] The server applies natural language processing (NLP) techniques to the collected review information. Using Python's NLTK and spaCy libraries, it tokenizes the review text, tags it with parts of speech, and performs dependency analysis to extract important keywords and phrases. For example, keywords such as "battery life" and "camera performance" are extracted.

[0521] 3. Review categorization and scoring

[0522] The server categorizes the reviews based on the analysis results and scores them as positive or negative based on their rating. This information is used to generate a comprehensive review summary, which includes the distribution of rating scores, representative examples of positive and negative comments, and the overall rating trend.

[0523] 4. Terminals and User Interfaces

[0524] In response to a user request, the device displays the review summary information received from the server on the user interface. The user refers to this summary to check the overall evaluation of a specific product or service and make a purchasing decision. As a specific example, the device screen may display the following message: "This smartphone has many positive reviews regarding its battery life, but there are 20 negative comments regarding its camera performance."

[0525] Prompt Sentence Examples

[0526] Below is a specific example of a request that a user sends to the server.

[0527] "Collect and analyze reviews of the latest smartphones and generate summaries."

[0528] "Please categorize the positive and negative reviews of a particular cosmetic product and give us a representative opinion for each."

[0529] The system of the present invention allows users to efficiently collect fragmented review information on the Internet and obtain highly reliable information based on the analysis results. This allows users to quickly and accurately understand product and service evaluations and make wise purchasing decisions. It also provides companies with an effective means of accurately understanding the evaluations of their own products and services.

[0530] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0531] Step 1:

[0532] A user sends a request to the server to collect review information about a particular product or service. The input is the product or service information specified by the user, including a prompt such as "Please collect reviews about the latest smartphones." The output is a response confirming that the server received the request.

[0533] Step 2:

[0534] The server launches a web crawler using the Python Scrapy library based on the request received from the user. The input is the product or service information in the user request and a list of websites to be collected. The output is log information indicating that the crawler has started accessing the websites to be collected.

[0535] Step 3:

[0536] The server's web crawler visits specified websites (e.g., major e-commerce sites or review sites) and collects review information (review text, rating score, reviewer information, and posting date and time). The input is the HTML data of the web pages accessed by the crawler. The output is a dataset of extracted review information, which is stored in a database such as MySQL or MongoDB.

[0537] Step 4:

[0538] The server applies natural language processing (NLP) to the database containing the collected review information. It uses Python's NLTK and spaCy libraries to tokenize the review text and perform part-of-speech tagging and dependency analysis. The input is the review text data in the database. The output is a dataset containing the analyzed tokens, part-of-speech tags, and dependencies.

[0539] Step 5:

[0540] The server extracts important keywords and phrases based on the analysis results and categorizes each review into a category. It also scores reviews as positive or negative based on their rating scores. The input is the dataset after NLP analysis. The output is a dataset of reviews categorized and scored.

[0541] Step 6:

[0542] The server generates a review summary based on the scored review information. This summary includes the distribution of rating scores, representative examples of positive and negative opinions, and overall rating trends. The input is the categorized and scored review information. The output is the generated review summary.

[0543] Step 7:

[0544] The device displays review summary information received from the server on the user interface in response to a user request. The input is the review summary sent from the server. The output is the review summary displayed on the user interface, which the user can refer to to check the evaluation of the product or service. Specifically, the device displays the message, "This smartphone has many positive reviews regarding its battery life, but there are 20 negative comments regarding its camera performance."

[0545] (Application example 1)

[0546] 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."

[0547] Conventional online review information collection and analysis systems have made it difficult for users to accurately and efficiently collect and understand review information about specific products. This requires manually examining a huge number of reviews, which is time-consuming and labor-intensive. Furthermore, there are limited means of providing reliable review information in a visually easy-to-understand format. Therefore, there is a need for technology that can more quickly and accurately assist users in making purchasing decisions.

[0548] 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.

[0549] In this invention, the server includes means for automatically collecting evaluation information on the Internet, means for analyzing the collected evaluation information and extracting key keywords and evaluation scores, means for generating an evaluation summary based on the analyzed information and displaying it through a user interface, input means for users to search for evaluation information on a specific product, and means for generating a summary based on the evaluation information and displaying it in a visually easy-to-understand format, thereby enabling users to quickly and efficiently obtain and understand highly reliable evaluation information on a specific product.

[0550] The "Internet" is a global network that interconnects computers and networks around the world, enabling the exchange of information.

[0551] "Evaluation information" refers to data such as users' opinions and impressions about products and services, and evaluation scores.

[0552] "Means" refers to methods, techniques, devices, etc. used to achieve a certain purpose.

[0553] "Collection" is the act of gathering information or data.

[0554] "Analysis" is the process of examining collected information and data in detail to clarify its structure, meaning, and relationships.

[0555] "Points" refers to particularly important elements or main points.

[0556] A "keyword" refers to an important word that characterizes a particular piece of information.

[0557] "Evaluation score" refers to data that quantifies the evaluation of a product or service.

[0558] "Summary" refers to a concise report or overview that summarises detailed information.

[0559] "User interface" refers to the means or environment through which a user and a system interact.

[0560] "Input" is the act of providing data or instructions to a system.

[0561] "Visual" refers to a form or expression that can be perceived by the eye.

[0562] A "website" is a collection of information hosted on the Internet.

[0563] A "web crawler" is a program that automatically crawls web pages on the Internet and collects information.

[0564] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[0565] "Positive" refers to positive opinions and evaluations.

[0566] "Negative" refers to negative opinions or evaluations.

[0567] A "trend" refers to a movement or change that shows a particular direction or tendency.

[0568] This invention is a system that automatically collects and analyzes evaluation information related to specific products and services from the Internet, and generates easy-to-understand summaries based on the results to support users' purchasing decisions. This system is mainly composed of a server and user terminals.

[0569] 1. System Configuration

[0570] Hardware

[0571] 1. Server

[0572] Cloud servers (e.g., AWS, GCP) are used to collect, analyze, and generate summaries of evaluation information.

[0573] 2. User Device

[0574] It is designed for smartphones and tablets and is responsible for displaying summary information.

[0575] software

[0576] 1. Web crawler

[0577] Using tools such as Scrapy, evaluation information is automatically collected from specific websites on the Internet.

[0578] 2. Natural Language Processing (NLP)

[0579] Using tools such as NLTK or Spacy, the collected evaluation information is analyzed to extract keywords and evaluation scores.

[0580] 3. Database

[0581] Use MySQL or MongoDB to store the collected and analyzed data.

[0582] 4. Front-end

[0583] It uses React Native to display summary information in a user-friendly interface.

[0584] 5. Backend

[0585] Use Django to coordinate data processing and user interface processing.

[0586] 2. System operation explanation

[0587] Collection Phase

[0588] A user submits a search request for rating information about a particular product.

[0589] The server receives the request and launches a web crawler, which then visits the specified website, collects rating information (e.g., review text, rating score, reviewer information, and posting date and time), and stores it in a database.

[0590] Analysis Phase

[0591] The server analyzes the rating information stored in the database using NLP techniques, including tokenization, part-of-speech tagging, and dependency analysis to extract important keywords and phrases.

[0592] Based on the extracted keywords and phrases, the rating information is categorized and positive and negative reviews are distinguished.

[0593] Summary Generation Phase

[0594] The server generates a summary of the review information based on the analysis results, including the distribution of review scores, representative examples of positive and negative comments, and overall review trends.

[0595] The user terminal displays the summary information through a user interface, allowing the user to quickly check the evaluation of the product or service.

[0596] 3. Examples and prompts

[0597] Examples:

[0598] Let's say a user wants to find out reviews about "XYZ Smartphone."

[0599] Users search for "XYZ smartphone" in the app, and the server collects, analyzes, and generates summaries of related reviews from various websites.

[0600] The analysis results in a conclusion such as, "This smartphone is rated for its long battery life, but there are 20 negative reviews about its camera performance."

[0601] Example prompt sentence:

[0602] Please analyze online reviews for XYZ smartphone and tell me the number of positive and negative reviews and what the representative opinions are.

[0603] The present invention allows users to quickly and efficiently obtain highly reliable evaluation information, which can support their own purchasing decisions.

[0604] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0605] Step 1:

[0606] User enters a search request

[0607] When a user wants to know evaluation information about a specific product, the user starts up an app on a device such as a smartphone or tablet and inputs the name of the product.

[0608] Input: Specific product name

[0609] Output: Search request

[0610] Step 2:

[0611] The device sends a request to the server

[0612] After the terminal accepts the user input, it sends the search request to the server.

[0613] Input: Search request

[0614] Output: Request sent to server

[0615] Step 3:

[0616] The server launches the web crawler

[0617] After the server receives the search request, it launches a web crawler to collect reputation information from the specified website.

[0618] Input: Search request

[0619] Output: Collected rating information (review text, rating score, reviewer information, posting date and time)

[0620] Step 4:

[0621] The server stores the rating information in a database

[0622] The server stores the collected evaluation information in a database.

[0623] Input: Collected evaluation information

[0624] Output: Evaluation information stored in a database

[0625] Step 5:

[0626] The server analyzes the evaluation information.

[0627] The server retrieves the rating information from the database and analyzes it using natural language processing techniques, including tokenization, part-of-speech tagging, and dependency analysis.

[0628] Input: Rating information retrieved from the database

[0629] Output: Analyzed data (keywords, rating scores, reviewer information, etc.)

[0630] Step 6:

[0631] The server categorizes positive and negative reviews

[0632] Based on the analyzed data, the server categorizes the rating information and distinguishes between positive and negative reviews.

[0633] Input: Parsed data

[0634] Output: Categorized positive and negative reviews

[0635] Step 7:

[0636] The server generates a summary of the rating information

[0637] The server generates a summary of the rating information based on representative examples of positive and negative reviews, the distribution of rating scores, and overall rating trends.

[0638] Input: Categorized positive and negative reviews

[0639] Output: Summary of evaluation information

[0640] Step 8:

[0641] User terminal displays summary information

[0642] When the user accesses the app again, it receives a summary of the rating information from the server and displays it visually through the user interface.

[0643] Input: Summary of rating information sent from the server

[0644] Output: A summary of the evaluation information displayed on the terminal.

[0645] 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.

[0646] System configuration

[0647] This invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. By combining this system with an emotion engine, it has the ability to recognize user emotions from review information and generate emotion scores. The following describes the system's configuration and processing in detail.

[0648] Detailed description of the program's processing

[0649] Collecting review information

[0650] The server receives a request from a user to collect review information about a particular product or service.

[0651] Based on the request, the server launches a web crawler, which then visits the specified website (e.g., a major e-commerce site, a review site, etc.) and automatically collects review information.

[0652] The server's crawler extracts the review text, rating score, reviewer information, and posting date and time from each web page, and stores this information in a database.

[0653] Review information analysis

[0654] The server applies natural language processing (NLP) algorithms to the collected review information, including tokenizing the review text, tagging parts of speech, dependency analysis, and keyword extraction, to identify important elements of the review.

[0655] The server categorizes each review based on the extracted keywords and phrases, and scores them based on the evaluation score, thereby classifying reviews into positive and negative.

[0656] Emotion recognition by emotion engine

[0657] The server uses an emotion engine to recognize the user's emotion from the analyzed review information. The emotion engine determines the emotion (e.g., joy, anger, sadness, etc.) of the user who wrote the review based on the context and keywords in the review text.

[0658] The server quantifies the emotions determined by the emotion engine and generates an emotion score, which is reflected in the review summary along with the positive / negative rating.

[0659] Generate and display summaries

[0660] The server generates a summary of the review information based on the analysis results and sentiment scores, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends, as well as the sentiment scores.

[0661] In response to a user request, the terminal displays the summary information received from the server on a user interface. The user can refer to this summary to check product and service ratings and make a purchasing decision.

[0662] Specific examples

[0663] 1. Collecting review information

[0664] A user sends a request to a server to collect review information about a particular smartphone.

[0665] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[0666] 2. Review information analysis

[0667] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[0668] The server scores each review based on its rating score and categorizes the reviews as positive and negative, for example identifying 60 out of 80 positive reviews about "battery life."

[0669] 3. Emotion Recognition by Emotion Engine

[0670] The server uses an emotion engine to analyze the review text and recognize the user's emotion. For example, it recognizes the emotion "joy" from a review text such as "very satisfied" and generates an emotion score.

[0671] The server stores the sentiment scores along with other analytical results in a database, recording the emotional state of each review.

[0672] 4. Generating and displaying summaries

[0673] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and sentiment scores and provides it to the user. This summary includes the distribution of rating scores, representative positive and negative opinions, overall rating trends, and sentiment scores.

[0674] The device displays this summary information on the user interface, allowing users to check product ratings. For example, users can learn that "this smartphone has a good battery life, but the camera performance has 20 negative reviews," and also get a conclusion that "user sentiment scores are high overall."

[0675] As described above, the system of the present invention efficiently collects and analyzes fragmented review information on the Internet and recognizes user sentiment to provide highly reliable information. This system allows users to quickly and accurately understand product and service ratings, and provides companies with an effective means of accurately understanding the ratings of their products and services.

[0676] The processing flow will be explained below.

[0677] Step 1: Submit a request to collect reviews

[0678] A user transmits a request for collecting review information about a specific product or service from a terminal, and the request includes identification information of the product or service (e.g., product ID or name).

[0679] Step 2: Launch the web crawler

[0680] The server receives a request from a user and launches a web crawler, which visits specific pre-defined websites and review sites (e.g., major e-commerce sites, word-of-mouth sites, etc.).

[0681] Step 3: Extracting review information

[0682] The server's crawler analyzes the specified web page and extracts review information, which means analyzing HTML tags and obtaining necessary data such as the review text, rating score, reviewer information, and posting date and time.

[0683] Step 4: Save your review information

[0684] The server stores the extracted review information in a database, adding the website information and product ID from which it was collected as metadata, making the information traceable later.

[0685] Step 5: Begin data analysis

[0686] The server initiates an analysis process of the collected review information based on periodic batch processing or user requests, a procedure for efficiently processing large amounts of data.

[0687] Step 6: Applying Natural Language Processing (NLP)

[0688] The server applies NLP algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, analyzing dependencies, and extracting keywords, thereby identifying important elements of the review.

[0689] Step 7: Categorizing and scoring reviews

[0690] The server categorizes each review based on keywords and phrases extracted through NLP processing, then distinguishes between positive and negative reviews and assigns them a score based on the review's rating.

[0691] Step 8: Emotion Recognition with the Emotion Engine

[0692] The server uses an emotion engine to recognize the user's emotion from the analyzed review information. The emotion engine determines the emotion (e.g., joy, anger, sadness, etc.) of the user who wrote the review based on the context and keywords in the review text.

[0693] The server quantifies the emotions determined by the emotion engine and generates an emotion score, which is reflected in the review summary along with the positive / negative rating.

[0694] Step 9: Generate a summary

[0695] The server generates a summary of the review information based on the analysis results and sentiment scores, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends, as well as the sentiment scores.

[0696] Step 10: Send summary information

[0697] The server sends the generated summary information to the user's terminal for provision to the user. This communication is performed in real time in response to a request or periodically as a batch process.

[0698] Step 11: View summary information

[0699] The terminal displays the received summary information on a user interface, allowing the user to quickly understand the evaluation of a particular product or service and make a purchasing decision.

[0700] As described above, this system efficiently collects and analyzes online review information and provides highly reliable information that also reflects user sentiment, allowing users to quickly and accurately understand product and service evaluations and make appropriate decisions.

[0701] Example 2

[0702] 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."

[0703] Currently, there is a huge amount of review information on the Internet, but this information is fragmented, and checking each review individually is extremely time-consuming. This makes it difficult for users to quickly and accurately grasp the overall evaluation of a product or service. Furthermore, reviews contain many emotional elements, which can significantly influence the evaluation. However, there is a lack of systems that can properly recognize and reflect emotions. This makes it difficult to provide reliable review summaries.

[0704] 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 a means for automatically collecting review information on the Internet, a means for analyzing the collected review information and extracting key keywords and evaluation scores, a means for recognizing the user's emotions based on the analyzed information and generating an emotion score, and a means for generating a summary of the review and displaying it through a user interface. This allows the user to easily grasp a reliable comprehensive evaluation that includes emotional elements without having to check each piece of fragmented information one by one.

[0705] The "Internet" is an information and communications infrastructure that interconnects computer networks around the world.

[0706] "Review information" is data such as sentences and scores written by users evaluating products or services.

[0707] "Means of collection" refers to the method of automatically obtaining information from the Internet using programs such as web crawlers.

[0708] "Means of analysis" refers to a method of using natural language processing technology to analyze collected data and extract important information.

[0709] "Keywords" refer to important words or phrases extracted from the review text.

[0710] A "rating score" is a numerical rating given by a user to a product or service.

[0711] The "emotion score" is an index that quantifies the emotions of the user who wrote the review by analyzing the review text.

[0712] A "summary" is information that aggregates and summarizes the analysis results of multiple reviews.

[0713] A "user interface" refers to the screen and operating means that allow a user to interact with a system.

[0714] A "crawler" is a program that automatically visits multiple web pages and collects information.

[0715] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[0716] This invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. By combining this system with an emotion engine, it has the ability to recognize user emotions from review information and generate emotion scores. The following describes the system's configuration and processing in detail.

[0717] System configuration

[0718] The system consists of three main components: a server, a terminal, and a user.

[0719] server

[0720] The server plays a key role in collecting and analyzing review information. The server has the following functions:

[0721] Collection function: Launches a web crawler and collects review information from specified e-commerce sites and review sites.

[0722] Analysis function: Using natural language processing (NLP) technology, collected review information is analyzed and keywords and rating scores are extracted.

[0723] Emotion Recognition: An emotion engine is used to recognize user emotions from the review text and generate an emotion score.

[0724] Summary generation function: Generates a summary of review information based on analysis results and sentiment scores.

[0725] Terminal

[0726] The terminal provides an interface that receives requests from users and displays summary information. The terminal has the following functions:

[0727] Request reception function: Receives collection requests from users and sends them to the server.

[0728] Summary display function: Displays the summary information received from the server on the user interface.

[0729] User

[0730] Users can check review information about products and services through their terminals and make purchasing decisions.

[0731] Hardware and software used

[0732] Web crawler: A program that crawls designated websites on the Internet to collect review information.

[0733] Natural language processing (NLP) technology: Technology that analyzes collected review information and extracts important keywords and phrases.

[0734] Sentiment engine: A software tool that recognizes user emotions from review text and generates an emotion score.

[0735] Database: A data storage system for storing the collected and analyzed review information and generated sentiment scores.

[0736] Specific examples

[0737] 1. Collecting review information

[0738] A user sends a request to a server to collect review information about a particular smartphone.

[0739] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[0740] 2. Review information analysis

[0741] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[0742] 3. Emotion Recognition by Emotion Engine

[0743] The server uses an emotion engine to analyze the review text and recognize the user's emotion. For example, it recognizes the emotion "joy" from a review text such as "very satisfied" and generates an emotion score.

[0744] 4. Generating and displaying summaries

[0745] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and sentiment scores and provides it to the user. This summary includes the distribution of rating scores, representative positive and negative opinions, overall rating trends, and sentiment scores.

[0746] The device displays this summary information on the user interface, allowing users to check product ratings. For example, users can learn that "this smartphone has a good battery life, but the camera performance has 20 negative reviews," and also get a conclusion that "user sentiment scores are high overall."

[0747] Through the above processing steps, the system provides users with reliable review information and analysis results.

[0748] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0749] Step 1: Accepting a user request

[0750] A user requests the collection of review information about a specific product or service from a terminal.

[0751] Input: User request (e.g., "Collect review information for smartphone A")

[0752] The terminal receives the user's request and transmits the request to the server.

[0753] Output: Request data to the server

[0754] Step 2: Collect review information

[0755] The server launches a web crawler based on a request received from a user.

[0756] Input: User request data

[0757] The server's web crawler patrols designated e-commerce sites and review sites and automatically collects relevant review information.

[0758] What happens: The server configures the web crawler and performs the task of gathering reviews for the specified sites.

[0759] Output: Collected review information (review text, rating score, reviewer information, posting date and time)

[0760] Step 3: Analyze review information

[0761] The server applies natural language processing (NLP) techniques to the review information stored in the database.

[0762] Input: Saved review information data

[0763] The server tokenizes (divides) the review text into words, tags it with parts of speech, and performs dependency analysis to extract important keywords and phrases.

[0764] How it works: The server uses an NLP library to tokenize and parse specific words in the review text, assigning them parts of speech, and also performs contextual analysis of the review text to extract important phrases.

[0765] Output: Parsed keywords, rating score, reviewer information

[0766] Step 4: Emotion Recognition with the Emotion Engine

[0767] The server recognizes the user's emotions from the analyzed review information using an emotion engine.

[0768] Input: Analyzed keywords, rating score, reviewer information

[0769] The server determines the sentiment based on the context and keywords in the review text and generates a numerical sentiment score.

[0770] Specific operation: The server uses an emotion recognition algorithm to perform contextual analysis of the review text and quantify the type of emotion (e.g., joy, anger) and its intensity.

[0771] Output: Sentiment score

[0772] Step 5: Generate a summary

[0773] The server generates a summary of the review information based on the analysis results and the sentiment score.

[0774] Input: Analysis results, emotion score

[0775] The server generates a summary that includes the distribution of rating scores, representative examples of positive and negative comments, overall rating trends, and sentiment scores.

[0776] What it does: The server combines the analysis results with sentiment scores, aggregates the data, and generates a visualizable summary. It also extracts representative opinions from each review and analyzes overall rating trends.

[0777] Output: Generated review summary

[0778] Step 6: View the summary

[0779] When a user requests summary information, the server sends the summary to the terminal.

[0780] Input: User summary information request

[0781] The server transmits the summary information generated in response to the request to the terminal.

[0782] The terminal displays the received summary information on a user interface.

[0783] Specific operation: The terminal analyzes the summary information received from the server and displays it visually in an easy-to-understand manner on the user interface.

[0784] Output: Review summary displayed to the user (e.g., distribution of individual rating scores, positive / negative opinions, sentiment score)

[0785] Through the above processing steps, the system provides users with reliable review information and analysis results.

[0786] (Application example 2)

[0787] 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."

[0788] Conventional methods for evaluating products based on online reviews have been difficult to utilize effectively due to the sheer volume of review information. Furthermore, simply averaging the evaluation scores can miss important information and users' feelings contained in the review text. In order for users to evaluate products and services and make appropriate purchasing decisions, it is necessary to efficiently analyze review information and provide reliable summary information.

[0789] 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 automatically collecting review information from the Internet, means for analyzing the collected review information and extracting key keywords and evaluation scores, means for generating review summaries based on the analyzed information and displaying them through a user interface, and means including an emotion recognition engine for generating emotion scores from the review text and reflecting the generated emotion scores in the summaries. This allows users to grasp not only the overall evaluation of the review information but also the emotional trends contained in each review, enabling reliable product evaluations based on the review information.

[0790] "Online review information" refers to user evaluations and opinions of products and services posted on the Internet.

[0791] "Automatic collection means" refers to a device or program that automatically collects designated information without requiring manual operation by the user.

[0792] "Analysis means" refers to the functions and algorithms used to decipher the collected data and understand its contents.

[0793] "Key keywords" refer to words or phrases that have particular significance in the review text.

[0794] "Rating score" refers to the numerical rating given by users in reviews of products or services.

[0795] "Means for generating summaries" refers to the function for creating summarized information based on collected and analyzed information.

[0796] "Means of displaying through a user interface" refers to the screens and operating methods used to visually present information to the user.

[0797] An "emotion recognition engine" refers to technology or algorithms that identify and quantify emotions in documents through text analysis.

[0798] An "emotion score" is a numerical value that quantitatively represents the emotion contained in a sentence.

[0799] A "generative AI model" refers to an artificial intelligence model built to learn from data and perform a specified task (in this case, sentiment analysis or review analysis).

[0800] The present invention is a system that automatically collects and analyzes review information available on the Internet and provides users with highly reliable review summaries. This system has the function of generating emotion scores from reviews using an emotion recognition engine and reflecting these emotion scores in summaries. The detailed configuration and operation of the system are described below.

[0801] System configuration

[0802] The system mainly consists of the following components:

[0803] 1. Server:

[0804] How to collect review information: The server launches a web crawler to automatically collect review information from specific websites. This crawls through specified web pages and extracts data such as the review text, rating score, reviewer information, and posting date and time.

[0805] Analysis method: We use natural language processing (NLP) technology to analyze the collected review information, specifically tokenization, part-of-speech tagging, dependency analysis, and keyword extraction.

[0806] Emotion Recognition Engine: Recognizes emotions from the review text and generates an emotion score. The emotion engine determines the user's emotion (e.g., joy, anger, sadness, etc.) based on the context and keywords in the review text.

[0807] 2. Terminal:

[0808] Summary generation method: The server analyzes the data and generates a sentiment score to generate a reliable review summary, which includes the distribution of rating scores, representative examples of positive and negative opinions, overall rating trends, and sentiment scores.

[0809] User interface: The user can view and refer to the review summary received from the server through the terminal.

[0810] Software and hardware used in the implementation

[0811] Hardware: Smartphones, tablets, PCs

[0812] software:

[0813] Python: a programming language

[0814] BeautifulSoup: A library for web crawling

[0815] requests: HTTP request library

[0816] nltk: A natural language processing library

[0817] SentimentIntensityAnalyzer: An nltk module for sentiment analysis.

[0818] Specific examples

[0819] If a user wants to check reviews of the latest smartphones, they first search for "latest smartphones" within the app. This search request is sent to a server, which collects review information from the specified online shopping site. The collected reviews are analyzed using natural language processing technology to extract important keywords and evaluation scores. An emotion recognition engine is also used to generate an emotion score from the review text. Finally, a reliable review summary is generated based on the analysis results and emotion score, and is displayed through the device's user interface.

[0820] Prompt Sentence Examples

[0821] "Collect reviews of this smartphone and display an overall rating summary based on the analysis results and sentiment score."

[0822] In this way, users can understand not only the overall evaluation of the review information but also the emotional trends contained in each review, enabling them to make highly reliable purchasing decisions.

[0823] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0824] Step 1:

[0825] A user sends a request to the server from their device to collect reviews about a specific product. The input is the user's request, including the product name and category. The server receives this request.

[0826] Step 2:

[0827] The server launches a web crawler and crawls designated websites to collect review information. The input is the user's request, and the output is the collected review information. Specifically, the server uses the crawler to analyze web pages and extract the review text, rating score, reviewer information, and posting date and time.

[0828] Step 3:

[0829] The server stores the collected review information in a database. The input is the collected review information, and the output is analyzable data stored in the database. The server stores the review information by dividing it into fields.

[0830] Step 4:

[0831] The server applies natural language processing (NLP) techniques to the stored review information. The input is the review information read from the database, and the output is the parsed review content. Specifically, the server tokenizes each review text and performs part-of-speech tagging, dependency analysis, keyword extraction, etc.

[0832] Step 5:

[0833] The server categorizes reviews based on the results of natural language processing and categorizes positive and negative reviews based on their rating scores. The input is the analyzed review content, and the output is the categorized reviews. The server also calculates the distribution of rating scores.

[0834] Step 6:

[0835] The server uses an emotion recognition engine to analyze the review text and generate an emotion score. The input is the categorized review content, and the output is an emotion score for each review. Specifically, the server calculates an emotion score (e.g., joy, anger, sadness, etc.) based on the context and keywords in the review text.

[0836] Step 7:

[0837] The server generates a review summary based on the analysis results and sentiment scores. The input is the review analysis results and sentiment scores, and the output is a review summary. The server creates a comprehensive summary that includes the distribution of rating scores, representative examples of positive and negative opinions, overall rating trends, and sentiment scores.

[0838] Step 8:

[0839] When a user requests summary information, the server generates a review summary and sends it to the terminal. The input is the user's summary request, and the output is the summary information.

[0840] Step 9:

[0841] The terminal displays the review summaries received from the server on the user interface. The input is the summary information sent from the server, and the output is the review summary displayed on the user interface. This allows the user to refer to the product ratings and make a purchasing decision.

[0842] 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.

[0843] 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.

[0844] 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.

[0845] [Third embodiment]

[0846] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0847] 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.

[0848] 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).

[0849] 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.

[0850] 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.

[0851] 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).

[0852] 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.

[0853] 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.

[0854] 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.

[0855] 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.

[0856] 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.

[0857] 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."

[0858] System configuration

[0859] The present invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. The configuration and processing of the system are described in detail below.

[0860] Detailed description of the program's processing

[0861] Collecting review information

[0862] The server receives a request from a user to collect review information about a particular product or service.

[0863] Based on the request, the server launches a web crawler, which then visits the specified website (e.g., a major e-commerce site, a review site, etc.) and automatically collects review information.

[0864] The server's crawler extracts the review text, rating score, reviewer information, and posting date and time from each web page, and stores this information in a database.

[0865] Review information analysis

[0866] The server applies natural language processing (NLP) algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, and performing dependency analysis to extract important keywords and phrases.

[0867] The server categorizes each review based on the extracted keywords and phrases, and scores them based on the evaluation score, thereby classifying reviews into positive and negative.

[0868] Generate and display summaries

[0869] The server generates a summary of the reviews based on the analysis, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends.

[0870] In response to a user request, the terminal displays the summary information received from the server on a user interface. The user can refer to this summary to check product and service ratings and make a purchasing decision.

[0871] Specific examples

[0872] 1. Collecting review information

[0873] A user sends a request to a server to collect review information about a particular smartphone.

[0874] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[0875] 2. Review information analysis

[0876] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[0877] The server scores each review based on its rating score and categorizes the reviews as positive and negative, for example identifying 60 out of 80 positive reviews about "battery life."

[0878] 3. Generating and displaying summaries

[0879] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and provides it to the user, including the distribution of rating scores, representative positive and negative opinions, and overall rating trends.

[0880] The device displays this summary information on the user interface, and the user can refer to it to check product reviews. For example, by looking at the summary information, the user can come to a conclusion such as "This smartphone is rated for its long battery life, but there are 20 negative reviews about its camera performance."

[0881] As described above, the system of the present invention can efficiently collect and analyze fragmented review information on the Internet and provide highly reliable information to support user decision-making. This system allows consumers to quickly and accurately understand product and service ratings, and provides companies with an effective means of accurately understanding the ratings of their products and services.

[0882] The processing flow will be explained below.

[0883] Step 1: Submit a request to collect reviews

[0884] A user transmits a request for collecting review information about a specific product or service from a terminal, and the request includes identification information of the product or service (e.g., product ID or name).

[0885] Step 2: Launch the web crawler

[0886] The server receives a request from a user and launches a web crawler, which visits specific pre-defined websites and review sites (e.g., major e-commerce sites, word-of-mouth sites, etc.).

[0887] Step 3: Extracting review information

[0888] The server's crawler analyzes the specified web page and extracts review information, which means analyzing HTML tags and obtaining necessary data such as the review text, rating score, reviewer information, and posting date and time.

[0889] Step 4: Save your review information

[0890] The server stores the extracted review information in a database, adding the website information and product ID from which it was collected as metadata, making the information traceable later.

[0891] Step 5: Begin data analysis

[0892] The server initiates an analysis process of the collected review information based on periodic batch processing or user requests, a procedure for efficiently processing large amounts of data.

[0893] Step 6: Applying Natural Language Processing (NLP)

[0894] The server applies NLP algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, analyzing dependencies, and extracting keywords, thereby identifying important elements of the review.

[0895] Step 7: Categorizing and scoring reviews

[0896] The server categorizes each review based on keywords and phrases extracted through NLP processing, then distinguishes between positive and negative reviews and assigns them a score based on the review's rating.

[0897] Step 8: Generate a summary

[0898] The server generates a summary of the review information based on the analysis results, including the distribution of rating scores, representative positive and negative opinions, and overall rating trends.

[0899] Step 9: Send summary information

[0900] The server sends the generated summary information to the user's terminal for provision to the user. This communication is performed in real time in response to a request or periodically as a batch process.

[0901] Step 10: View summary information

[0902] The terminal displays the received summary information on a user interface, allowing the user to quickly understand the evaluation of a particular product or service and make a purchasing decision.

[0903] The above is the specific program processing flow of this system, which enables users to efficiently obtain highly reliable review information and make decisions.

[0904] Example 1

[0905] 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."

[0906] Until now, much of the review information on the Internet has been scattered individually, making it difficult for users to efficiently compare and analyze it. Additionally, grasping the reliability and overall trends of collected review information has been a time-consuming and labor-intensive process. Furthermore, the lack of a means to appropriately utilize natural language processing technology when analyzing review information has made it difficult to create accurate review summaries. Therefore, there is a need for a system that can efficiently collect and analyze review information and provide highly reliable review summaries.

[0907] 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.

[0908] In this invention, the server includes: means for receiving a request from a user to collect review information about a specific product or service; means for automatically collecting review information from the Internet; means for analyzing the collected review information, tokenizing the review text using natural language processing technology, performing part-of-speech tagging and dependency analysis, and extracting key keywords and evaluation scores; means for generating a summary of the review based on the analyzed information and displaying the summary through a user interface, including the distribution of evaluation scores, representative positive and negative opinions, and overall evaluation trends; and means for displaying the summary information on the user interface in response to a user request. This allows fragmented review information to be efficiently collected and analyzed, enabling users to easily check reviews and make quick and accurate purchasing decisions for products and services.

[0909] The "means for receiving a request from a user to collect review information about a specific product or service" refers to a function in which a user sends a request to collect reviews about a specific product or service to a server and the request is received.

[0910] "Means for automatically collecting review information on the Internet" refers to a function for automatically obtaining review information from specific websites on the Internet, including the use of a crawler program for this purpose.

[0911] "Natural language processing technology" is a technology for analyzing review text, and performs processes such as tokenization, part-of-speech tagging, and dependency analysis.

[0912] "Tokenizing the review text" refers to the process of dividing the entire review text into words and phrases.

[0913] "Part-of-speech tagging" refers to the process of assigning each token the appropriate part of speech (noun, verb, adjective, etc.).

[0914] "Dependency parsing" refers to a parsing technique that reveals the relationships between words in a sentence (e.g., subject and verb, modifier and noun, etc.).

[0915] "Means for extracting key keywords and evaluation scores" refers to the function of extracting important words and phrases from the review text, as well as user evaluations.

[0916] "Generating review summaries" refers to the process of creating information summarizing overall ratings, positive and negative opinions, etc., based on collected and analyzed review information.

[0917] "Means for displaying through a user interface" refers to a function for visually displaying the generated review summary on the user's screen.

[0918] "Rating score distribution" refers to data showing how the rating scores of collected reviews are distributed.

[0919] "Representative positive and negative opinions" refers to the most common positive and negative comments found across all reviews.

[0920] "Overall rating trends" refers to data showing changes in review content and rating scores over a certain period of time.

[0921] A "crawler" is a program that automatically crawls specific websites on the Internet and collects specified information.

[0922] The present invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. The main feature of this system is that it automatically collects review information based on user requests, analyzes it using natural language processing technology, and displays it as a summary. Specific embodiments are described below.

[0923] System configuration

[0924] This system is composed of three main components: a server, a terminal, and a user.

[0925] 1. Server

[0926] The server receives a request from a user to collect review information about a specific product or service. Upon receiving the request, the server launches a web crawler program using the Python Scrapy library to collect review information from the specified website. The collected review information (e.g., review text, rating score, reviewer information, and posting date and time) is stored in a database such as MySQL or MongoDB.

[0927] 2. Review information analysis

[0928] The server applies natural language processing (NLP) techniques to the collected review information. Using Python's NLTK and spaCy libraries, it tokenizes the review text, tags it with parts of speech, and performs dependency analysis to extract important keywords and phrases. For example, keywords such as "battery life" and "camera performance" are extracted.

[0929] 3. Review categorization and scoring

[0930] The server categorizes the reviews based on the analysis results and scores them as positive or negative based on their rating. This information is used to generate a comprehensive review summary, which includes the distribution of rating scores, representative examples of positive and negative comments, and the overall rating trend.

[0931] 4. Terminals and User Interfaces

[0932] In response to a user request, the device displays the review summary information received from the server on the user interface. The user refers to this summary to check the overall evaluation of a specific product or service and make a purchasing decision. As a specific example, the device screen may display the following message: "This smartphone has many positive reviews regarding its battery life, but there are 20 negative comments regarding its camera performance."

[0933] Prompt Sentence Examples

[0934] Below is a specific example of a request that a user sends to the server.

[0935] "Collect and analyze reviews of the latest smartphones and generate summaries."

[0936] "Please categorize the positive and negative reviews of a particular cosmetic product and give us a representative opinion for each."

[0937] The system of the present invention allows users to efficiently collect fragmented review information on the Internet and obtain highly reliable information based on the analysis results. This allows users to quickly and accurately understand product and service evaluations and make wise purchasing decisions. It also provides companies with an effective means of accurately understanding the evaluations of their own products and services.

[0938] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0939] Step 1:

[0940] A user sends a request to the server to collect review information about a particular product or service. The input is the product or service information specified by the user, including a prompt such as "Please collect reviews about the latest smartphones." The output is a response confirming that the server received the request.

[0941] Step 2:

[0942] The server launches a web crawler using the Python Scrapy library based on the request received from the user. The input is the product or service information in the user request and a list of websites to be collected. The output is log information indicating that the crawler has started accessing the websites to be collected.

[0943] Step 3:

[0944] The server's web crawler visits specified websites (e.g., major e-commerce sites or review sites) and collects review information (review text, rating score, reviewer information, and posting date and time). The input is the HTML data of the web pages accessed by the crawler. The output is a dataset of extracted review information, which is stored in a database such as MySQL or MongoDB.

[0945] Step 4:

[0946] The server applies natural language processing (NLP) to the database containing the collected review information. It uses Python's NLTK and spaCy libraries to tokenize the review text and perform part-of-speech tagging and dependency analysis. The input is the review text data in the database. The output is a dataset containing the analyzed tokens, part-of-speech tags, and dependencies.

[0947] Step 5:

[0948] The server extracts important keywords and phrases based on the analysis results and categorizes each review into a category. It also scores reviews as positive or negative based on their rating scores. The input is the dataset after NLP analysis. The output is a dataset of reviews categorized and scored.

[0949] Step 6:

[0950] The server generates a review summary based on the scored review information. This summary includes the distribution of rating scores, representative examples of positive and negative opinions, and overall rating trends. The input is the categorized and scored review information. The output is the generated review summary.

[0951] Step 7:

[0952] The device displays review summary information received from the server on the user interface in response to a user request. The input is the review summary sent from the server. The output is the review summary displayed on the user interface, which the user can refer to to check the evaluation of the product or service. Specifically, the device displays the message, "This smartphone has many positive reviews regarding its battery life, but there are 20 negative comments regarding its camera performance."

[0953] (Application example 1)

[0954] 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."

[0955] Conventional online review information collection and analysis systems have made it difficult for users to accurately and efficiently collect and understand review information about specific products. This requires manually examining a huge number of reviews, which is time-consuming and labor-intensive. Furthermore, there are limited means of providing reliable review information in a visually easy-to-understand format. Therefore, there is a need for technology that can more quickly and accurately assist users in making purchasing decisions.

[0956] 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.

[0957] In this invention, the server includes means for automatically collecting evaluation information on the Internet, means for analyzing the collected evaluation information and extracting key keywords and evaluation scores, means for generating an evaluation summary based on the analyzed information and displaying it through a user interface, input means for users to search for evaluation information on a specific product, and means for generating a summary based on the evaluation information and displaying it in a visually easy-to-understand format, thereby enabling users to quickly and efficiently obtain and understand highly reliable evaluation information on a specific product.

[0958] The "Internet" is a global network that interconnects computers and networks around the world, enabling the exchange of information.

[0959] "Evaluation information" refers to data such as users' opinions and impressions about products and services, and evaluation scores.

[0960] "Means" refers to methods, techniques, devices, etc. used to achieve a certain purpose.

[0961] "Collection" is the act of gathering information or data.

[0962] "Analysis" is the process of examining collected information and data in detail to clarify its structure, meaning, and relationships.

[0963] "Points" refers to particularly important elements or main points.

[0964] A "keyword" refers to an important word that characterizes a particular piece of information.

[0965] "Evaluation score" refers to data that quantifies the evaluation of a product or service.

[0966] "Summary" refers to a concise report or overview that summarises detailed information.

[0967] "User interface" refers to the means or environment through which a user and a system interact.

[0968] "Input" is the act of providing data or instructions to a system.

[0969] "Visual" refers to a form or expression that can be perceived by the eye.

[0970] A "website" is a collection of information hosted on the Internet.

[0971] A "web crawler" is a program that automatically crawls web pages on the Internet and collects information.

[0972] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[0973] "Positive" refers to positive opinions and evaluations.

[0974] "Negative" refers to negative opinions or evaluations.

[0975] A "trend" refers to a movement or change that shows a particular direction or tendency.

[0976] This invention is a system that automatically collects and analyzes evaluation information related to specific products and services from the Internet, and generates easy-to-understand summaries based on the results to support users' purchasing decisions. This system is mainly composed of a server and user terminals.

[0977] 1. System Configuration

[0978] Hardware

[0979] 1. Server

[0980] Cloud servers (e.g., AWS, GCP) are used to collect, analyze, and generate summaries of evaluation information.

[0981] 2. User Device

[0982] It is designed for smartphones and tablets and is responsible for displaying summary information.

[0983] software

[0984] 1. Web crawler

[0985] Using tools such as Scrapy, evaluation information is automatically collected from specific websites on the Internet.

[0986] 2. Natural Language Processing (NLP)

[0987] Using tools such as NLTK or Spacy, the collected evaluation information is analyzed to extract keywords and evaluation scores.

[0988] 3. Database

[0989] Use MySQL or MongoDB to store the collected and analyzed data.

[0990] 4. Front-end

[0991] It uses React Native to display summary information in a user-friendly interface.

[0992] 5. Backend

[0993] Use Django to coordinate data processing and user interface processing.

[0994] 2. System operation explanation

[0995] Collection Phase

[0996] A user submits a search request for rating information about a particular product.

[0997] The server receives the request and launches a web crawler, which then visits the specified website, collects rating information (e.g., review text, rating score, reviewer information, and posting date and time), and stores it in a database.

[0998] Analysis Phase

[0999] The server analyzes the rating information stored in the database using NLP techniques, including tokenization, part-of-speech tagging, and dependency analysis to extract important keywords and phrases.

[1000] Based on the extracted keywords and phrases, the rating information is categorized and positive and negative reviews are distinguished.

[1001] Summary Generation Phase

[1002] The server generates a summary of the review information based on the analysis results, including the distribution of review scores, representative examples of positive and negative comments, and overall review trends.

[1003] The user terminal displays the summary information through a user interface, allowing the user to quickly check the evaluation of the product or service.

[1004] 3. Examples and prompts

[1005] Examples:

[1006] Let's say a user wants to find out reviews about "XYZ Smartphone."

[1007] Users search for "XYZ smartphone" in the app, and the server collects, analyzes, and generates summaries of related reviews from various websites.

[1008] The analysis results in a conclusion such as, "This smartphone is rated for its long battery life, but there are 20 negative reviews about its camera performance."

[1009] Example prompt sentence:

[1010] Please analyze online reviews for XYZ smartphone and tell me the number of positive and negative reviews and what the representative opinions are.

[1011] The present invention allows users to quickly and efficiently obtain highly reliable evaluation information, which can support their own purchasing decisions.

[1012] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1013] Step 1:

[1014] User enters a search request

[1015] When a user wants to know evaluation information about a specific product, the user starts up an app on a device such as a smartphone or tablet and inputs the name of the product.

[1016] Input: Specific product name

[1017] Output: Search request

[1018] Step 2:

[1019] The device sends a request to the server

[1020] After the terminal accepts the user input, it sends the search request to the server.

[1021] Input: Search request

[1022] Output: Request sent to server

[1023] Step 3:

[1024] The server launches the web crawler

[1025] After the server receives the search request, it launches a web crawler to collect reputation information from the specified website.

[1026] Input: Search request

[1027] Output: Collected rating information (review text, rating score, reviewer information, posting date and time)

[1028] Step 4:

[1029] The server stores the rating information in a database

[1030] The server stores the collected evaluation information in a database.

[1031] Input: Collected evaluation information

[1032] Output: Evaluation information stored in a database

[1033] Step 5:

[1034] The server analyzes the evaluation information.

[1035] The server retrieves the rating information from the database and analyzes it using natural language processing techniques, including tokenization, part-of-speech tagging, and dependency analysis.

[1036] Input: Rating information retrieved from the database

[1037] Output: Analyzed data (keywords, rating scores, reviewer information, etc.)

[1038] Step 6:

[1039] The server categorizes positive and negative reviews

[1040] Based on the analyzed data, the server categorizes the rating information and distinguishes between positive and negative reviews.

[1041] Input: Parsed data

[1042] Output: Categorized positive and negative reviews

[1043] Step 7:

[1044] The server generates a summary of the rating information

[1045] The server generates a summary of the rating information based on representative examples of positive and negative reviews, the distribution of rating scores, and overall rating trends.

[1046] Input: Categorized positive and negative reviews

[1047] Output: Summary of evaluation information

[1048] Step 8:

[1049] User terminal displays summary information

[1050] When the user accesses the app again, it receives a summary of the rating information from the server and displays it visually through the user interface.

[1051] Input: Summary of rating information sent from the server

[1052] Output: A summary of the evaluation information displayed on the terminal.

[1053] 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.

[1054] System configuration

[1055] This invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. By combining this system with an emotion engine, it has the ability to recognize user emotions from review information and generate emotion scores. The following describes the system's configuration and processing in detail.

[1056] Detailed description of the program's processing

[1057] Collecting review information

[1058] The server receives a request from a user to collect review information about a particular product or service.

[1059] Based on the request, the server launches a web crawler, which then visits the specified website (e.g., a major e-commerce site, a review site, etc.) and automatically collects review information.

[1060] The server's crawler extracts the review text, rating score, reviewer information, and posting date and time from each web page, and stores this information in a database.

[1061] Review information analysis

[1062] The server applies natural language processing (NLP) algorithms to the collected review information, including tokenizing the review text, tagging parts of speech, dependency analysis, and keyword extraction, to identify important elements of the review.

[1063] The server categorizes each review based on the extracted keywords and phrases, and scores them based on the evaluation score, thereby classifying reviews into positive and negative.

[1064] Emotion recognition by emotion engine

[1065] The server uses an emotion engine to recognize the user's emotion from the analyzed review information. The emotion engine determines the emotion (e.g., joy, anger, sadness, etc.) of the user who wrote the review based on the context and keywords in the review text.

[1066] The server quantifies the emotions determined by the emotion engine and generates an emotion score, which is reflected in the review summary along with the positive / negative rating.

[1067] Generate and display summaries

[1068] The server generates a summary of the review information based on the analysis results and sentiment scores, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends, as well as the sentiment scores.

[1069] In response to a user request, the terminal displays the summary information received from the server on a user interface. The user can refer to this summary to check product and service ratings and make a purchasing decision.

[1070] Specific examples

[1071] 1. Collecting review information

[1072] A user sends a request to a server to collect review information about a particular smartphone.

[1073] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[1074] 2. Review information analysis

[1075] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[1076] The server scores each review based on its rating score and categorizes the reviews as positive and negative, for example identifying 60 out of 80 positive reviews about "battery life."

[1077] 3. Emotion Recognition by Emotion Engine

[1078] The server uses an emotion engine to analyze the review text and recognize the user's emotion. For example, it recognizes the emotion "joy" from a review text such as "very satisfied" and generates an emotion score.

[1079] The server stores the sentiment scores along with other analytical results in a database, recording the emotional state of each review.

[1080] 4. Generating and displaying summaries

[1081] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and sentiment scores and provides it to the user. This summary includes the distribution of rating scores, representative positive and negative opinions, overall rating trends, and sentiment scores.

[1082] The device displays this summary information on the user interface, allowing users to check product ratings. For example, users can learn that "this smartphone has a good battery life, but the camera performance has 20 negative reviews," and also get a conclusion that "user sentiment scores are high overall."

[1083] As described above, the system of the present invention efficiently collects and analyzes fragmented review information on the Internet and recognizes user sentiment to provide highly reliable information. This system allows users to quickly and accurately understand product and service ratings, and provides companies with an effective means of accurately understanding the ratings of their products and services.

[1084] The processing flow will be explained below.

[1085] Step 1: Submit a request to collect reviews

[1086] A user transmits a request for collecting review information about a specific product or service from a terminal, and the request includes identification information of the product or service (e.g., product ID or name).

[1087] Step 2: Launch the web crawler

[1088] The server receives a request from a user and launches a web crawler, which visits specific pre-defined websites and review sites (e.g., major e-commerce sites, word-of-mouth sites, etc.).

[1089] Step 3: Extracting review information

[1090] The server's crawler analyzes the specified web page and extracts review information, which means analyzing HTML tags and obtaining necessary data such as the review text, rating score, reviewer information, and posting date and time.

[1091] Step 4: Save your review information

[1092] The server stores the extracted review information in a database, adding the website information and product ID from which it was collected as metadata, making the information traceable later.

[1093] Step 5: Begin data analysis

[1094] The server initiates an analysis process of the collected review information based on periodic batch processing or user requests, a procedure for efficiently processing large amounts of data.

[1095] Step 6: Applying Natural Language Processing (NLP)

[1096] The server applies NLP algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, analyzing dependencies, and extracting keywords, thereby identifying important elements of the review.

[1097] Step 7: Categorizing and scoring reviews

[1098] The server categorizes each review based on keywords and phrases extracted through NLP processing, then distinguishes between positive and negative reviews and assigns them a score based on the review's rating.

[1099] Step 8: Emotion Recognition with the Emotion Engine

[1100] The server uses an emotion engine to recognize the user's emotion from the analyzed review information. The emotion engine determines the emotion (e.g., joy, anger, sadness, etc.) of the user who wrote the review based on the context and keywords in the review text.

[1101] The server quantifies the emotions determined by the emotion engine and generates an emotion score, which is reflected in the review summary along with the positive / negative rating.

[1102] Step 9: Generate a summary

[1103] The server generates a summary of the review information based on the analysis results and sentiment scores, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends, as well as the sentiment scores.

[1104] Step 10: Send summary information

[1105] The server sends the generated summary information to the user's terminal for provision to the user. This communication is performed in real time in response to a request or periodically as a batch process.

[1106] Step 11: View summary information

[1107] The terminal displays the received summary information on a user interface, allowing the user to quickly understand the evaluation of a particular product or service and make a purchasing decision.

[1108] As described above, this system efficiently collects and analyzes online review information and provides highly reliable information that also reflects user sentiment, allowing users to quickly and accurately understand product and service evaluations and make appropriate decisions.

[1109] Example 2

[1110] 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."

[1111] Currently, there is a huge amount of review information on the Internet, but this information is fragmented, and checking each review individually is extremely time-consuming. This makes it difficult for users to quickly and accurately grasp the overall evaluation of a product or service. Furthermore, reviews contain many emotional elements, which can significantly influence the evaluation. However, there is a lack of systems that can properly recognize and reflect emotions. This makes it difficult to provide reliable review summaries.

[1112] 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 a means for automatically collecting review information on the Internet, a means for analyzing the collected review information and extracting key keywords and evaluation scores, a means for recognizing the user's emotions based on the analyzed information and generating an emotion score, and a means for generating a summary of the review and displaying it through a user interface. This allows the user to easily grasp a reliable comprehensive evaluation that includes emotional elements without having to check each piece of fragmented information one by one.

[1113] The "Internet" is an information and communications infrastructure that interconnects computer networks around the world.

[1114] "Review information" is data such as sentences and scores written by users evaluating products or services.

[1115] "Means of collection" refers to the method of automatically obtaining information from the Internet using programs such as web crawlers.

[1116] "Means of analysis" refers to a method of using natural language processing technology to analyze collected data and extract important information.

[1117] "Keywords" refer to important words or phrases extracted from the review text.

[1118] A "rating score" is a numerical rating given by a user to a product or service.

[1119] The "emotion score" is an index that quantifies the emotions of the user who wrote the review by analyzing the review text.

[1120] A "summary" is information that aggregates and summarizes the analysis results of multiple reviews.

[1121] A "user interface" refers to the screen and operating means that allow a user to interact with a system.

[1122] A "crawler" is a program that automatically visits multiple web pages and collects information.

[1123] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[1124] This invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. By combining this system with an emotion engine, it has the ability to recognize user emotions from review information and generate emotion scores. The following describes the system's configuration and processing in detail.

[1125] System configuration

[1126] The system consists of three main components: a server, a terminal, and a user.

[1127] server

[1128] The server plays a key role in collecting and analyzing review information. The server has the following functions:

[1129] Collection function: Launches a web crawler and collects review information from specified e-commerce sites and review sites.

[1130] Analysis function: Using natural language processing (NLP) technology, collected review information is analyzed and keywords and rating scores are extracted.

[1131] Emotion Recognition: An emotion engine is used to recognize user emotions from the review text and generate an emotion score.

[1132] Summary generation function: Generates a summary of review information based on analysis results and sentiment scores.

[1133] Terminal

[1134] The terminal provides an interface that receives requests from users and displays summary information. The terminal has the following functions:

[1135] Request reception function: Receives collection requests from users and sends them to the server.

[1136] Summary display function: Displays the summary information received from the server on the user interface.

[1137] User

[1138] Users can check review information about products and services through their terminals and make purchasing decisions.

[1139] Hardware and software used

[1140] Web crawler: A program that crawls designated websites on the Internet to collect review information.

[1141] Natural language processing (NLP) technology: Technology that analyzes collected review information and extracts important keywords and phrases.

[1142] Sentiment engine: A software tool that recognizes user emotions from review text and generates an emotion score.

[1143] Database: A data storage system for storing the collected and analyzed review information and generated sentiment scores.

[1144] Specific examples

[1145] 1. Collecting review information

[1146] A user sends a request to a server to collect review information about a particular smartphone.

[1147] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[1148] 2. Review information analysis

[1149] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[1150] 3. Emotion Recognition by Emotion Engine

[1151] The server uses an emotion engine to analyze the review text and recognize the user's emotion. For example, it recognizes the emotion "joy" from a review text such as "very satisfied" and generates an emotion score.

[1152] 4. Generating and displaying summaries

[1153] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and sentiment scores and provides it to the user. This summary includes the distribution of rating scores, representative positive and negative opinions, overall rating trends, and sentiment scores.

[1154] The device displays this summary information on the user interface, allowing users to check product ratings. For example, users can learn that "this smartphone has a good battery life, but the camera performance has 20 negative reviews," and also get a conclusion that "user sentiment scores are high overall."

[1155] Through the above processing steps, the system provides users with reliable review information and analysis results.

[1156] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1157] Step 1: Accepting a user request

[1158] A user requests the collection of review information about a specific product or service from a terminal.

[1159] Input: User request (e.g., "Collect review information for smartphone A")

[1160] The terminal receives the user's request and transmits the request to the server.

[1161] Output: Request data to the server

[1162] Step 2: Collect review information

[1163] The server launches a web crawler based on a request received from a user.

[1164] Input: User request data

[1165] The server's web crawler patrols designated e-commerce sites and review sites and automatically collects relevant review information.

[1166] What happens: The server configures the web crawler and performs the task of gathering reviews for the specified sites.

[1167] Output: Collected review information (review text, rating score, reviewer information, posting date and time)

[1168] Step 3: Analyze review information

[1169] The server applies natural language processing (NLP) techniques to the review information stored in the database.

[1170] Input: Saved review information data

[1171] The server tokenizes (divides) the review text into words, tags it with parts of speech, and performs dependency analysis to extract important keywords and phrases.

[1172] How it works: The server uses an NLP library to tokenize and parse specific words in the review text, assigning them parts of speech, and also performs contextual analysis of the review text to extract important phrases.

[1173] Output: Parsed keywords, rating score, reviewer information

[1174] Step 4: Emotion Recognition with the Emotion Engine

[1175] The server recognizes the user's emotions from the analyzed review information using an emotion engine.

[1176] Input: Analyzed keywords, rating score, reviewer information

[1177] The server determines the sentiment based on the context and keywords in the review text and generates a numerical sentiment score.

[1178] Specific operation: The server uses an emotion recognition algorithm to perform contextual analysis of the review text and quantify the type of emotion (e.g., joy, anger) and its intensity.

[1179] Output: Sentiment score

[1180] Step 5: Generate a summary

[1181] The server generates a summary of the review information based on the analysis results and the sentiment score.

[1182] Input: Analysis results, emotion score

[1183] The server generates a summary that includes the distribution of rating scores, representative examples of positive and negative comments, overall rating trends, and sentiment scores.

[1184] What it does: The server combines the analysis results with sentiment scores, aggregates the data, and generates a visualizable summary. It also extracts representative opinions from each review and analyzes overall rating trends.

[1185] Output: Generated review summary

[1186] Step 6: View the summary

[1187] When a user requests summary information, the server sends the summary to the terminal.

[1188] Input: User summary information request

[1189] The server transmits the summary information generated in response to the request to the terminal.

[1190] The terminal displays the received summary information on a user interface.

[1191] Specific operation: The terminal analyzes the summary information received from the server and displays it visually in an easy-to-understand manner on the user interface.

[1192] Output: Review summary displayed to the user (e.g., distribution of individual rating scores, positive / negative opinions, sentiment score)

[1193] Through the above processing steps, the system provides users with reliable review information and analysis results.

[1194] (Application example 2)

[1195] 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."

[1196] Conventional methods for evaluating products based on online reviews have been difficult to utilize effectively due to the sheer volume of review information. Furthermore, simply averaging the evaluation scores can miss important information and users' feelings contained in the review text. In order for users to evaluate products and services and make appropriate purchasing decisions, it is necessary to efficiently analyze review information and provide reliable summary information.

[1197] 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 automatically collecting review information from the Internet, means for analyzing the collected review information and extracting key keywords and evaluation scores, means for generating review summaries based on the analyzed information and displaying them through a user interface, and means including an emotion recognition engine for generating emotion scores from the review text and reflecting the generated emotion scores in the summaries. This allows users to grasp not only the overall evaluation of the review information but also the emotional trends contained in each review, enabling reliable product evaluations based on the review information.

[1198] "Online review information" refers to user evaluations and opinions of products and services posted on the Internet.

[1199] "Automatic collection means" refers to a device or program that automatically collects designated information without requiring manual operation by the user.

[1200] "Analysis means" refers to the functions and algorithms used to decipher the collected data and understand its contents.

[1201] "Key keywords" refer to words or phrases that have particular significance in the review text.

[1202] "Rating score" refers to the numerical rating given by users in reviews of products or services.

[1203] "Means for generating summaries" refers to the function for creating summarized information based on collected and analyzed information.

[1204] "Means of displaying through a user interface" refers to the screens and operating methods used to visually present information to the user.

[1205] An "emotion recognition engine" refers to technology or algorithms that identify and quantify emotions in documents through text analysis.

[1206] An "emotion score" is a numerical value that quantitatively represents the emotion contained in a sentence.

[1207] A "generative AI model" refers to an artificial intelligence model built to learn from data and perform a specified task (in this case, sentiment analysis or review analysis).

[1208] The present invention is a system that automatically collects and analyzes review information available on the Internet and provides users with highly reliable review summaries. This system has the function of generating emotion scores from reviews using an emotion recognition engine and reflecting these emotion scores in summaries. The detailed configuration and operation of the system are described below.

[1209] System configuration

[1210] The system mainly consists of the following components:

[1211] 1. Server:

[1212] How to collect review information: The server launches a web crawler to automatically collect review information from specific websites. This crawls through specified web pages and extracts data such as the review text, rating score, reviewer information, and posting date and time.

[1213] Analysis method: We use natural language processing (NLP) technology to analyze the collected review information, specifically tokenization, part-of-speech tagging, dependency analysis, and keyword extraction.

[1214] Emotion Recognition Engine: Recognizes emotions from the review text and generates an emotion score. The emotion engine determines the user's emotion (e.g., joy, anger, sadness, etc.) based on the context and keywords in the review text.

[1215] 2. Terminal:

[1216] Summary generation method: The server analyzes the data and generates a sentiment score to generate a reliable review summary, which includes the distribution of rating scores, representative examples of positive and negative opinions, overall rating trends, and sentiment scores.

[1217] User interface: The user can view and refer to the review summary received from the server through the terminal.

[1218] Software and hardware used in the implementation

[1219] Hardware: Smartphones, tablets, PCs

[1220] software:

[1221] Python: a programming language

[1222] BeautifulSoup: A library for web crawling

[1223] requests: HTTP request library

[1224] nltk: A natural language processing library

[1225] SentimentIntensityAnalyzer: An nltk module for sentiment analysis.

[1226] Specific examples

[1227] If a user wants to check reviews of the latest smartphones, they first search for "latest smartphones" within the app. This search request is sent to a server, which collects review information from the specified online shopping site. The collected reviews are analyzed using natural language processing technology to extract important keywords and evaluation scores. An emotion recognition engine is also used to generate an emotion score from the review text. Finally, a reliable review summary is generated based on the analysis results and emotion score, and is displayed through the device's user interface.

[1228] Prompt Sentence Examples

[1229] "Collect reviews of this smartphone and display an overall rating summary based on the analysis results and sentiment score."

[1230] In this way, users can understand not only the overall evaluation of the review information but also the emotional trends contained in each review, enabling them to make highly reliable purchasing decisions.

[1231] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1232] Step 1:

[1233] A user sends a request to the server from their device to collect reviews about a specific product. The input is the user's request, including the product name and category. The server receives this request.

[1234] Step 2:

[1235] The server launches a web crawler and crawls designated websites to collect review information. The input is the user's request, and the output is the collected review information. Specifically, the server uses the crawler to analyze web pages and extract the review text, rating score, reviewer information, and posting date and time.

[1236] Step 3:

[1237] The server stores the collected review information in a database. The input is the collected review information, and the output is analyzable data stored in the database. The server stores the review information by dividing it into fields.

[1238] Step 4:

[1239] The server applies natural language processing (NLP) techniques to the stored review information. The input is the review information read from the database, and the output is the parsed review content. Specifically, the server tokenizes each review text and performs part-of-speech tagging, dependency analysis, keyword extraction, etc.

[1240] Step 5:

[1241] The server categorizes reviews based on the results of natural language processing and categorizes positive and negative reviews based on their rating scores. The input is the analyzed review content, and the output is the categorized reviews. The server also calculates the distribution of rating scores.

[1242] Step 6:

[1243] The server uses an emotion recognition engine to analyze the review text and generate an emotion score. The input is the categorized review content, and the output is an emotion score for each review. Specifically, the server calculates an emotion score (e.g., joy, anger, sadness, etc.) based on the context and keywords in the review text.

[1244] Step 7:

[1245] The server generates a review summary based on the analysis results and sentiment scores. The input is the review analysis results and sentiment scores, and the output is a review summary. The server creates a comprehensive summary that includes the distribution of rating scores, representative examples of positive and negative opinions, overall rating trends, and sentiment scores.

[1246] Step 8:

[1247] When a user requests summary information, the server generates a review summary and sends it to the terminal. The input is the user's summary request, and the output is the summary information.

[1248] Step 9:

[1249] The terminal displays the review summaries received from the server on the user interface. The input is the summary information sent from the server, and the output is the review summary displayed on the user interface. This allows the user to refer to the product ratings and make a purchasing decision.

[1250] 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.

[1251] 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.

[1252] 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.

[1253] [Fourth embodiment]

[1254] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1255] 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.

[1256] 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).

[1257] 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.

[1258] 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.

[1259] 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).

[1260] 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.

[1261] 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.

[1262] 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.

[1263] 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.

[1264] 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.

[1265] 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.

[1266] 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."

[1267] System configuration

[1268] The present invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. The configuration and processing of the system are described in detail below.

[1269] Detailed description of the program's processing

[1270] Collecting review information

[1271] The server receives a request from a user to collect review information about a particular product or service.

[1272] Based on the request, the server launches a web crawler, which then visits the specified website (e.g., a major e-commerce site, a review site, etc.) and automatically collects review information.

[1273] The server's crawler extracts the review text, rating score, reviewer information, and posting date and time from each web page, and stores this information in a database.

[1274] Review information analysis

[1275] The server applies natural language processing (NLP) algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, and performing dependency analysis to extract important keywords and phrases.

[1276] The server categorizes each review based on the extracted keywords and phrases, and scores them based on the evaluation score, thereby classifying reviews into positive and negative.

[1277] Generate and display summaries

[1278] The server generates a summary of the reviews based on the analysis, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends.

[1279] In response to a user request, the terminal displays the summary information received from the server on a user interface. The user can refer to this summary to check product and service ratings and make a purchasing decision.

[1280] Specific examples

[1281] 1. Collecting review information

[1282] A user sends a request to a server to collect review information about a particular smartphone.

[1283] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[1284] 2. Review information analysis

[1285] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[1286] The server scores each review based on its rating score and categorizes the reviews as positive and negative, for example identifying 60 out of 80 positive reviews about "battery life."

[1287] 3. Generating and displaying summaries

[1288] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and provides it to the user, including the distribution of rating scores, representative positive and negative opinions, and overall rating trends.

[1289] The device displays this summary information on the user interface, and the user can refer to it to check product reviews. For example, by looking at the summary information, the user can come to a conclusion such as "This smartphone is rated for its long battery life, but there are 20 negative reviews about its camera performance."

[1290] As described above, the system of the present invention can efficiently collect and analyze fragmented review information on the Internet and provide highly reliable information to support user decision-making. This system allows consumers to quickly and accurately understand product and service ratings, and provides companies with an effective means of accurately understanding the ratings of their products and services.

[1291] The processing flow will be explained below.

[1292] Step 1: Submit a request to collect reviews

[1293] A user transmits a request for collecting review information about a specific product or service from a terminal, and the request includes identification information of the product or service (e.g., product ID or name).

[1294] Step 2: Launch the web crawler

[1295] The server receives a request from a user and launches a web crawler, which visits specific pre-defined websites and review sites (e.g., major e-commerce sites, word-of-mouth sites, etc.).

[1296] Step 3: Extracting review information

[1297] The server's crawler analyzes the specified web page and extracts review information, which means analyzing HTML tags and obtaining necessary data such as the review text, rating score, reviewer information, and posting date and time.

[1298] Step 4: Save your review information

[1299] The server stores the extracted review information in a database, adding the website information and product ID from which it was collected as metadata, making the information traceable later.

[1300] Step 5: Begin data analysis

[1301] The server initiates an analysis process of the collected review information based on periodic batch processing or user requests, a procedure for efficiently processing large amounts of data.

[1302] Step 6: Applying Natural Language Processing (NLP)

[1303] The server applies NLP algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, analyzing dependencies, and extracting keywords, thereby identifying important elements of the review.

[1304] Step 7: Categorizing and scoring reviews

[1305] The server categorizes each review based on keywords and phrases extracted through NLP processing, then distinguishes between positive and negative reviews and assigns them a score based on the review's rating.

[1306] Step 8: Generate a summary

[1307] The server generates a summary of the review information based on the analysis results, including the distribution of rating scores, representative positive and negative opinions, and overall rating trends.

[1308] Step 9: Send summary information

[1309] The server then sends the generated summary information to the user's terminal for provision to the user. This communication can be performed in real time in response to a request or periodically as a batch process.

[1310] Step 10: View summary information

[1311] The terminal displays the received summary information on a user interface, allowing the user to quickly understand the evaluation of a particular product or service and make a purchasing decision.

[1312] The above is the specific program processing flow of this system, which enables users to efficiently obtain highly reliable review information and make decisions.

[1313] Example 1

[1314] 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."

[1315] Until now, much of the review information on the Internet has been scattered individually, making it difficult for users to efficiently compare and analyze it. Additionally, grasping the reliability and overall trends of collected review information has been a time-consuming and labor-intensive process. Furthermore, the lack of a means to appropriately utilize natural language processing technology when analyzing review information has made it difficult to create accurate review summaries. Therefore, there is a need for a system that can efficiently collect and analyze review information and provide highly reliable review summaries.

[1316] 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.

[1317] In this invention, the server includes: means for receiving a request from a user to collect review information about a specific product or service; means for automatically collecting review information from the Internet; means for analyzing the collected review information, tokenizing the review text using natural language processing technology, performing part-of-speech tagging and dependency analysis, and extracting key keywords and evaluation scores; means for generating a summary of the review based on the analyzed information and displaying the summary through a user interface, including the distribution of evaluation scores, representative positive and negative opinions, and overall evaluation trends; and means for displaying the summary information on the user interface in response to a user request. This allows fragmented review information to be efficiently collected and analyzed, enabling users to easily check reviews and make quick and accurate purchasing decisions for products and services.

[1318] The "means for receiving a request from a user to collect review information about a specific product or service" refers to a function in which a user sends a request to collect reviews about a specific product or service to a server and the request is received.

[1319] "Means for automatically collecting review information on the Internet" refers to a function for automatically obtaining review information from specific websites on the Internet, including the use of a crawler program for this purpose.

[1320] "Natural language processing technology" is a technology for analyzing review text, and performs processes such as tokenization, part-of-speech tagging, and dependency analysis.

[1321] "Tokenizing the review text" refers to the process of dividing the entire review text into words and phrases.

[1322] "Part-of-speech tagging" refers to the process of assigning each token the appropriate part of speech (noun, verb, adjective, etc.).

[1323] "Dependency parsing" refers to a parsing technique that reveals the relationships between words in a sentence (e.g., subject and verb, modifier and noun, etc.).

[1324] "Means for extracting key keywords and evaluation scores" refers to the function of extracting important words and phrases from the review text, as well as user evaluations.

[1325] "Generating review summaries" refers to the process of creating information summarizing overall ratings, positive and negative opinions, etc., based on collected and analyzed review information.

[1326] "Means for displaying through a user interface" refers to a function for visually displaying the generated review summary on the user's screen.

[1327] "Rating score distribution" refers to data showing how the rating scores of collected reviews are distributed.

[1328] "Representative positive and negative opinions" refers to the most common positive and negative comments found across all reviews.

[1329] "Overall rating trends" refers to data showing changes in review content and rating scores over a certain period of time.

[1330] A "crawler" is a program that automatically crawls specific websites on the Internet and collects specified information.

[1331] The present invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. The main feature of this system is that it automatically collects review information based on user requests, analyzes it using natural language processing technology, and displays it as a summary. Specific embodiments are described below.

[1332] System configuration

[1333] This system is composed of three main components: a server, a terminal, and a user.

[1334] 1. Server

[1335] The server receives a request from a user to collect review information about a specific product or service. Upon receiving the request, the server launches a web crawler program using the Python Scrapy library to collect review information from the specified website. The collected review information (e.g., review text, rating score, reviewer information, and posting date and time) is stored in a database such as MySQL or MongoDB.

[1336] 2. Review information analysis

[1337] The server applies natural language processing (NLP) techniques to the collected review information. Using Python's NLTK and spaCy libraries, it tokenizes the review text, tags it with parts of speech, and performs dependency analysis to extract important keywords and phrases. For example, keywords such as "battery life" and "camera performance" are extracted.

[1338] 3. Review categorization and scoring

[1339] The server categorizes the reviews based on the analysis results and scores them as positive or negative based on their rating. This information is used to generate a comprehensive review summary, which includes the distribution of rating scores, representative examples of positive and negative comments, and the overall rating trend.

[1340] 4. Terminals and User Interfaces

[1341] In response to a user request, the device displays the review summary information received from the server on the user interface. The user refers to this summary to check the overall evaluation of a specific product or service and make a purchasing decision. As a specific example, the device screen may display the following message: "This smartphone has many positive reviews regarding its battery life, but there are 20 negative comments regarding its camera performance."

[1342] Prompt Sentence Examples

[1343] Below is a specific example of a request that a user sends to the server.

[1344] "Collect and analyze reviews of the latest smartphones and generate summaries."

[1345] "Please categorize the positive and negative reviews of a particular cosmetic product and give us a representative opinion for each."

[1346] The system of the present invention allows users to efficiently collect fragmented review information on the Internet and obtain highly reliable information based on the analysis results. This allows users to quickly and accurately understand product and service evaluations and make wise purchasing decisions. It also provides companies with an effective means of accurately understanding the evaluations of their own products and services.

[1347] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1348] Step 1:

[1349] A user sends a request to the server to collect review information about a particular product or service. The input is the product or service information specified by the user, including a prompt such as "Please collect reviews about the latest smartphones." The output is a response confirming that the server received the request.

[1350] Step 2:

[1351] The server launches a web crawler using the Python Scrapy library based on the request received from the user. The input is the product or service information in the user request and a list of websites to be collected. The output is log information indicating that the crawler has started accessing the websites to be collected.

[1352] Step 3:

[1353] The server's web crawler visits specified websites (e.g., major e-commerce sites or review sites) and collects review information (review text, rating score, reviewer information, and posting date and time). The input is the HTML data of the web pages accessed by the crawler. The output is a dataset of extracted review information, which is stored in a database such as MySQL or MongoDB.

[1354] Step 4:

[1355] The server applies natural language processing (NLP) to the database containing the collected review information. It uses Python's NLTK and spaCy libraries to tokenize the review text and perform part-of-speech tagging and dependency analysis. The input is the review text data in the database. The output is a dataset containing the analyzed tokens, part-of-speech tags, and dependencies.

[1356] Step 5:

[1357] The server extracts important keywords and phrases based on the analysis results and categorizes each review into a category. It also scores reviews as positive or negative based on their rating scores. The input is the dataset after NLP analysis. The output is a dataset of reviews categorized and scored.

[1358] Step 6:

[1359] The server generates a review summary based on the scored review information. This summary includes the distribution of rating scores, representative examples of positive and negative opinions, and overall rating trends. The input is the categorized and scored review information. The output is the generated review summary.

[1360] Step 7:

[1361] The device displays review summary information received from the server on the user interface in response to a user request. The input is the review summary sent from the server. The output is the review summary displayed on the user interface, which the user can refer to to check the evaluation of the product or service. Specifically, the device displays the message, "This smartphone has many positive reviews regarding its battery life, but there are 20 negative comments regarding its camera performance."

[1362] (Application example 1)

[1363] 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."

[1364] Conventional online review information collection and analysis systems have made it difficult for users to accurately and efficiently collect and understand review information about specific products. This requires manually examining a huge number of reviews, which is time-consuming and labor-intensive. Furthermore, there are limited means of providing reliable review information in a visually easy-to-understand format. Therefore, there is a need for technology that can more quickly and accurately assist users in making purchasing decisions.

[1365] 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.

[1366] In this invention, the server includes means for automatically collecting evaluation information on the Internet, means for analyzing the collected evaluation information and extracting key keywords and evaluation scores, means for generating an evaluation summary based on the analyzed information and displaying it through a user interface, input means for users to search for evaluation information on a specific product, and means for generating a summary based on the evaluation information and displaying it in a visually easy-to-understand format, thereby enabling users to quickly and efficiently obtain and understand highly reliable evaluation information on a specific product.

[1367] The "Internet" is a global network that interconnects computers and networks around the world, enabling the exchange of information.

[1368] "Evaluation information" refers to data such as users' opinions and impressions about products and services, and evaluation scores.

[1369] "Means" refers to methods, techniques, devices, etc. used to achieve a certain purpose.

[1370] "Collection" is the act of gathering information or data.

[1371] "Analysis" is the process of examining collected information and data in detail to clarify its structure, meaning, and relationships.

[1372] "Points" refers to particularly important elements or main points.

[1373] A "keyword" refers to an important word that characterizes a particular piece of information.

[1374] "Evaluation score" refers to data that quantifies the evaluation of a product or service.

[1375] "Summary" refers to a concise report or overview that summarises detailed information.

[1376] "User interface" refers to the means or environment through which a user and a system interact.

[1377] "Input" is the act of providing data or instructions to a system.

[1378] "Visual" refers to a form or expression that can be perceived by the eye.

[1379] A "website" is a collection of information hosted on the Internet.

[1380] A "web crawler" is a program that automatically crawls web pages on the Internet and collects information.

[1381] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[1382] "Positive" refers to positive opinions and evaluations.

[1383] "Negative" refers to negative opinions or evaluations.

[1384] A "trend" refers to a movement or change that shows a particular direction or tendency.

[1385] This invention is a system that automatically collects and analyzes evaluation information related to specific products and services from the Internet, and generates easy-to-understand summaries based on the results to support users' purchasing decisions. This system is mainly composed of a server and user terminals.

[1386] 1. System Configuration

[1387] Hardware

[1388] 1. Server

[1389] Cloud servers (e.g., AWS, GCP) are used to collect, analyze, and generate summaries of evaluation information.

[1390] 2. User Device

[1391] It is designed for smartphones and tablets and is responsible for displaying summary information.

[1392] software

[1393] 1. Web crawler

[1394] Using tools such as Scrapy, evaluation information is automatically collected from specific websites on the Internet.

[1395] 2. Natural Language Processing (NLP)

[1396] Using tools such as NLTK or Spacy, the collected evaluation information is analyzed to extract keywords and evaluation scores.

[1397] 3. Database

[1398] Use MySQL or MongoDB to store the collected and analyzed data.

[1399] 4. Front-end

[1400] It uses React Native to display summary information in a user-friendly interface.

[1401] 5. Backend

[1402] Use Django to coordinate data processing and user interface processing.

[1403] 2. System operation explanation

[1404] Collection Phase

[1405] A user submits a search request for rating information about a particular product.

[1406] The server receives the request and launches a web crawler, which then visits the specified website, collects rating information (e.g., review text, rating score, reviewer information, and posting date and time), and stores it in a database.

[1407] Analysis Phase

[1408] The server analyzes the rating information stored in the database using NLP techniques, including tokenization, part-of-speech tagging, and dependency analysis to extract important keywords and phrases.

[1409] Based on the extracted keywords and phrases, the rating information is categorized and positive and negative reviews are distinguished.

[1410] Summary Generation Phase

[1411] The server generates a summary of the review information based on the analysis results, including the distribution of review scores, representative examples of positive and negative comments, and overall review trends.

[1412] The user terminal displays the summary information through a user interface, allowing the user to quickly check the evaluation of the product or service.

[1413] 3. Examples and prompts

[1414] Examples:

[1415] Let's say a user wants to find out reviews about "XYZ Smartphone."

[1416] Users search for "XYZ smartphone" in the app, and the server collects, analyzes, and generates summaries of related reviews from various websites.

[1417] The analysis results in a conclusion such as, "This smartphone is rated for its long battery life, but there are 20 negative reviews about its camera performance."

[1418] Example prompt sentence:

[1419] Please analyze online reviews for XYZ smartphone and tell me the number of positive and negative reviews and what the representative opinions are.

[1420] The present invention allows users to quickly and efficiently obtain highly reliable evaluation information, which can support their own purchasing decisions.

[1421] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1422] Step 1:

[1423] User enters a search request

[1424] When a user wants to know evaluation information about a specific product, the user starts up an app on a device such as a smartphone or tablet and inputs the name of the product.

[1425] Input: Specific product name

[1426] Output: Search request

[1427] Step 2:

[1428] The device sends a request to the server

[1429] After the terminal accepts the user input, it sends the search request to the server.

[1430] Input: Search request

[1431] Output: Request sent to server

[1432] Step 3:

[1433] The server launches the web crawler

[1434] After the server receives the search request, it launches a web crawler to collect reputation information from the specified website.

[1435] Input: Search request

[1436] Output: Collected rating information (review text, rating score, reviewer information, posting date and time)

[1437] Step 4:

[1438] The server stores the rating information in a database

[1439] The server stores the collected evaluation information in a database.

[1440] Input: Collected evaluation information

[1441] Output: Evaluation information stored in a database

[1442] Step 5:

[1443] The server analyzes the evaluation information.

[1444] The server retrieves the rating information from the database and analyzes it using natural language processing techniques, including tokenization, part-of-speech tagging, and dependency analysis.

[1445] Input: Rating information retrieved from the database

[1446] Output: Analyzed data (keywords, rating scores, reviewer information, etc.)

[1447] Step 6:

[1448] The server categorizes positive and negative reviews

[1449] Based on the analyzed data, the server categorizes the rating information and distinguishes between positive and negative reviews.

[1450] Input: Parsed data

[1451] Output: Categorized positive and negative reviews

[1452] Step 7:

[1453] The server generates a summary of the rating information

[1454] The server generates a summary of the rating information based on representative examples of positive and negative reviews, the distribution of rating scores, and overall rating trends.

[1455] Input: Categorized positive and negative reviews

[1456] Output: Summary of evaluation information

[1457] Step 8:

[1458] User terminal displays summary information

[1459] When the user accesses the app again, it receives a summary of the rating information from the server and displays it visually through the user interface.

[1460] Input: Summary of rating information sent from the server

[1461] Output: A summary of the evaluation information displayed on the terminal.

[1462] 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.

[1463] System configuration

[1464] This invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. By combining this system with an emotion engine, it has the ability to recognize user emotions from review information and generate emotion scores. The following describes the system's configuration and processing in detail.

[1465] Detailed description of the program's processing

[1466] Collecting review information

[1467] The server receives a request from a user to collect review information about a particular product or service.

[1468] Based on the request, the server launches a web crawler, which then visits the specified website (e.g., a major e-commerce site, a review site, etc.) and automatically collects review information.

[1469] The server's crawler extracts the review text, rating score, reviewer information, and posting date and time from each web page, and stores this information in a database.

[1470] Review information analysis

[1471] The server applies natural language processing (NLP) algorithms to the collected review information, including tokenizing the review text, tagging parts of speech, dependency analysis, and keyword extraction, to identify important elements of the review.

[1472] The server categorizes each review based on the extracted keywords and phrases, and scores them based on the evaluation score, thereby classifying reviews into positive and negative.

[1473] Emotion recognition by emotion engine

[1474] The server uses an emotion engine to recognize the user's emotion from the analyzed review information. The emotion engine determines the emotion (e.g., joy, anger, sadness, etc.) of the user who wrote the review based on the context and keywords in the review text.

[1475] The server quantifies the emotions determined by the emotion engine and generates an emotion score, which is reflected in the review summary along with the positive / negative rating.

[1476] Generate and display summaries

[1477] The server generates a summary of the review information based on the analysis results and sentiment scores, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends, as well as the sentiment scores.

[1478] In response to a user request, the terminal displays the summary information received from the server on a user interface. The user can refer to this summary to check product and service ratings and make a purchasing decision.

[1479] Specific examples

[1480] 1. Collecting review information

[1481] A user sends a request to a server to collect review information about a particular smartphone.

[1482] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[1483] 2. Review information analysis

[1484] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[1485] The server scores each review based on its rating score and categorizes the reviews as positive and negative, for example identifying 60 out of 80 positive reviews about "battery life."

[1486] 3. Emotion Recognition by Emotion Engine

[1487] The server uses an emotion engine to analyze the review text and recognize the user's emotion. For example, it recognizes the emotion "joy" from a review text such as "very satisfied" and generates an emotion score.

[1488] The server stores the sentiment scores along with other analytical results in a database, recording the emotional state of each review.

[1489] 4. Generating and displaying summaries

[1490] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and sentiment scores and provides it to the user. This summary includes the distribution of rating scores, representative positive and negative opinions, overall rating trends, and sentiment scores.

[1491] The device displays this summary information on the user interface, allowing users to check product ratings. For example, users can learn that "this smartphone has a good battery life, but the camera performance has 20 negative reviews," and also get a conclusion that "user sentiment scores are high overall."

[1492] As described above, the system of the present invention efficiently collects and analyzes fragmented review information on the Internet and recognizes user sentiment to provide highly reliable information. This system allows users to quickly and accurately understand product and service ratings, and provides companies with an effective means of accurately understanding the ratings of their products and services.

[1493] The processing flow will be explained below.

[1494] Step 1: Submit a request to collect reviews

[1495] A user transmits a request for collecting review information about a specific product or service from a terminal, and the request includes identification information of the product or service (e.g., product ID or name).

[1496] Step 2: Launch the web crawler

[1497] The server receives a request from a user and launches a web crawler, which visits specific pre-defined websites and review sites (e.g., major e-commerce sites, word-of-mouth sites, etc.).

[1498] Step 3: Extracting review information

[1499] The server's crawler analyzes the specified web page and extracts review information, which means analyzing HTML tags and obtaining necessary data such as the review text, rating score, reviewer information, and posting date and time.

[1500] Step 4: Save your review information

[1501] The server stores the extracted review information in a database, adding the website information and product ID from which it was collected as metadata, making the information traceable later.

[1502] Step 5: Begin data analysis

[1503] The server initiates an analysis process of the collected review information based on periodic batch processing or user requests, a procedure for efficiently processing large amounts of data.

[1504] Step 6: Applying Natural Language Processing (NLP)

[1505] The server applies NLP algorithms to the collected review information, specifically tokenizing the review text, tagging it with parts of speech, analyzing dependencies, and extracting keywords, thereby identifying important elements of the review.

[1506] Step 7: Categorizing and scoring reviews

[1507] The server categorizes each review based on keywords and phrases extracted through NLP processing, then distinguishes between positive and negative reviews and assigns them a score based on the review's rating.

[1508] Step 8: Emotion Recognition with the Emotion Engine

[1509] The server uses an emotion engine to recognize the user's emotion from the analyzed review information. The emotion engine determines the emotion (e.g., joy, anger, sadness, etc.) of the user who wrote the review based on the context and keywords in the review text.

[1510] The server quantifies the emotions determined by the emotion engine and generates an emotion score, which is reflected in the review summary along with the positive / negative rating.

[1511] Step 9: Generate a summary

[1512] The server generates a summary of the review information based on the analysis results and sentiment scores, including the distribution of rating scores, representative examples of positive and negative comments, and overall rating trends, as well as the sentiment scores.

[1513] Step 10: Send summary information

[1514] The server sends the generated summary information to the user's terminal for provision to the user. This communication is performed in real time in response to a request or periodically as a batch process.

[1515] Step 11: View summary information

[1516] The terminal displays the received summary information on a user interface, allowing the user to quickly understand the evaluation of a particular product or service and make a purchasing decision.

[1517] As described above, this system efficiently collects and analyzes online review information and provides highly reliable information that also reflects user sentiment, allowing users to quickly and accurately understand product and service evaluations and make appropriate decisions.

[1518] Example 2

[1519] 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."

[1520] Currently, there is a huge amount of review information on the Internet, but this information is fragmented, and checking each review individually is extremely time-consuming. This makes it difficult for users to quickly and accurately grasp the overall evaluation of a product or service. Furthermore, reviews contain many emotional elements, which can significantly influence the evaluation. However, there is a lack of systems that can properly recognize and reflect emotions. This makes it difficult to provide reliable review summaries.

[1521] 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 a means for automatically collecting review information on the Internet, a means for analyzing the collected review information and extracting key keywords and evaluation scores, a means for recognizing the user's emotions based on the analyzed information and generating an emotion score, and a means for generating a summary of the review and displaying it through a user interface. This allows the user to easily grasp a reliable comprehensive evaluation that includes emotional elements without having to check each piece of fragmented information one by one.

[1522] The "Internet" is an information and communications infrastructure that interconnects computer networks around the world.

[1523] "Review information" is data such as sentences and scores written by users evaluating products or services.

[1524] "Means of collection" refers to the method of automatically obtaining information from the Internet using programs such as web crawlers.

[1525] "Means of analysis" refers to a method of using natural language processing technology to analyze collected data and extract important information.

[1526] "Keywords" refer to important words or phrases extracted from the review text.

[1527] A "rating score" is a numerical rating given by a user to a product or service.

[1528] The "emotion score" is an index that quantifies the emotions of the user who wrote the review by analyzing the review text.

[1529] A "summary" is information that aggregates and summarizes the analysis results of multiple reviews.

[1530] A "user interface" refers to the screen and operating means that allow a user to interact with a system.

[1531] A "crawler" is a program that automatically visits multiple web pages and collects information.

[1532] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[1533] This invention is a system that automatically collects and analyzes review information on the Internet and provides users with highly reliable review summaries. By combining this system with an emotion engine, it has the ability to recognize user emotions from review information and generate emotion scores. The following describes the system's configuration and processing in detail.

[1534] System configuration

[1535] The system consists of three main components: a server, a terminal, and a user.

[1536] server

[1537] The server plays a key role in collecting and analyzing review information. The server has the following functions:

[1538] Collection function: Launches a web crawler and collects review information from specified e-commerce sites and review sites.

[1539] Analysis function: Using natural language processing (NLP) technology, collected review information is analyzed and keywords and rating scores are extracted.

[1540] Emotion Recognition: An emotion engine is used to recognize user emotions from the review text and generate an emotion score.

[1541] Summary generation function: Generates a summary of review information based on analysis results and sentiment scores.

[1542] Terminal

[1543] The terminal provides an interface that receives requests from users and displays summary information. The terminal has the following functions:

[1544] Request reception function: Receives collection requests from users and sends them to the server.

[1545] Summary display function: Displays the summary information received from the server on the user interface.

[1546] User

[1547] Users can check review information about products and services through their terminals and make purchasing decisions.

[1548] Hardware and software used

[1549] Web crawler: A program that crawls designated websites on the Internet to collect review information.

[1550] Natural language processing (NLP) technology: Technology that analyzes collected review information and extracts important keywords and phrases.

[1551] Sentiment engine: A software tool that recognizes user emotions from review text and generates an emotion score.

[1552] Database: A data storage system for storing the collected and analyzed review information and generated sentiment scores.

[1553] Specific examples

[1554] 1. Collecting review information

[1555] A user sends a request to a server to collect review information about a particular smartphone.

[1556] The server launches a web crawler to collect smartphone review information from major e-commerce and review sites. The crawler extracts the review text, rating score, reviewer information, and posting date and time from the page and stores them in a database.

[1557] 2. Review information analysis

[1558] The server analyzes the collected smartphone review information using natural language processing technology, extracting key keywords such as "battery life" and "camera performance," and categorizing the reviews.

[1559] 3. Emotion Recognition by Emotion Engine

[1560] The server uses an emotion engine to analyze the review text and recognize the user's emotion. For example, it recognizes the emotion "joy" from a review text such as "very satisfied" and generates an emotion score.

[1561] 4. Generating and displaying summaries

[1562] When a user requests summary information on their smartphone, the server generates a summary based on the analysis results and sentiment scores and provides it to the user. This summary includes the distribution of rating scores, representative positive and negative opinions, overall rating trends, and sentiment scores.

[1563] The device displays this summary information on the user interface, allowing users to check product ratings. For example, users can learn that "this smartphone has a good battery life, but the camera performance has 20 negative reviews," and also get a conclusion that "user sentiment scores are high overall."

[1564] Through the above processing steps, the system provides users with reliable review information and analysis results.

[1565] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1566] Step 1: Accepting a user request

[1567] A user requests the collection of review information about a specific product or service from a terminal.

[1568] Input: User request (e.g., "Collect review information for smartphone A")

[1569] The terminal receives the user's request and transmits the request to the server.

[1570] Output: Request data to the server

[1571] Step 2: Collect review information

[1572] The server launches a web crawler based on a request received from a user.

[1573] Input: User request data

[1574] The server's web crawler patrols designated e-commerce sites and review sites and automatically collects relevant review information.

[1575] What happens: The server configures the web crawler and performs the task of gathering reviews for the specified sites.

[1576] Output: Collected review information (review text, rating score, reviewer information, posting date and time)

[1577] Step 3: Analyze review information

[1578] The server applies natural language processing (NLP) techniques to the review information stored in the database.

[1579] Input: Saved review information data

[1580] The server tokenizes (divides) the review text into words, tags it with parts of speech, and performs dependency analysis to extract important keywords and phrases.

[1581] How it works: The server uses an NLP library to tokenize and parse specific words in the review text, assigning them parts of speech, and also performs contextual analysis of the review text to extract important phrases.

[1582] Output: Parsed keywords, rating score, reviewer information

[1583] Step 4: Emotion Recognition with the Emotion Engine

[1584] The server recognizes the user's emotions from the analyzed review information using an emotion engine.

[1585] Input: Analyzed keywords, rating score, reviewer information

[1586] The server determines the sentiment based on the context and keywords in the review text and generates a numerical sentiment score.

[1587] Specific operation: The server uses an emotion recognition algorithm to perform contextual analysis of the review text and quantify the type of emotion (e.g., joy, anger) and its intensity.

[1588] Output: Sentiment score

[1589] Step 5: Generate a summary

[1590] The server generates a summary of the review information based on the analysis results and the sentiment score.

[1591] Input: Analysis results, emotion score

[1592] The server generates a summary that includes the distribution of rating scores, representative examples of positive and negative comments, overall rating trends, and sentiment scores.

[1593] What it does: The server combines the analysis results with sentiment scores, aggregates the data, and generates a visualizable summary. It also extracts representative opinions from each review and analyzes overall rating trends.

[1594] Output: Generated review summary

[1595] Step 6: View the summary

[1596] When a user requests summary information, the server sends the summary to the terminal.

[1597] Input: User summary information request

[1598] The server transmits the summary information generated in response to the request to the terminal.

[1599] The terminal displays the received summary information on a user interface.

[1600] Specific operation: The terminal analyzes the summary information received from the server and displays it visually in an easy-to-understand manner on the user interface.

[1601] Output: Review summary displayed to the user (e.g., distribution of individual rating scores, positive / negative opinions, sentiment score)

[1602] Through the above processing steps, the system provides users with reliable review information and analysis results.

[1603] (Application example 2)

[1604] 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."

[1605] Conventional methods for evaluating products based on online reviews have been difficult to utilize effectively due to the sheer volume of review information. Furthermore, simply averaging the evaluation scores can miss important information and users' feelings contained in the review text. In order for users to evaluate products and services and make appropriate purchasing decisions, it is necessary to efficiently analyze review information and provide reliable summary information.

[1606] 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 automatically collecting review information from the Internet, means for analyzing the collected review information and extracting key keywords and evaluation scores, means for generating review summaries based on the analyzed information and displaying them through a user interface, and means including an emotion recognition engine for generating emotion scores from the review text and reflecting the generated emotion scores in the summaries. This allows users to grasp not only the overall evaluation of the review information but also the emotional trends contained in each review, enabling reliable product evaluations based on the review information.

[1607] "Online review information" refers to user evaluations and opinions of products and services posted on the Internet.

[1608] "Automatic collection means" refers to a device or program that automatically collects designated information without requiring manual operation by the user.

[1609] "Analysis means" refers to the functions and algorithms used to decipher the collected data and understand its contents.

[1610] "Key keywords" refer to words or phrases that have particular significance in the review text.

[1611] "Rating score" refers to the numerical rating given by users in reviews of products or services.

[1612] "Means for generating summaries" refers to the function for creating summarized information based on collected and analyzed information.

[1613] "Means of displaying through a user interface" refers to the screens and operating methods used to visually present information to the user.

[1614] An "emotion recognition engine" refers to technology or algorithms that identify and quantify emotions in documents through text analysis.

[1615] An "emotion score" is a numerical value that quantitatively represents the emotion contained in a sentence.

[1616] A "generative AI model" refers to an artificial intelligence model built to learn from data and perform a specified task (in this case, sentiment analysis or review analysis).

[1617] The present invention is a system that automatically collects and analyzes review information available on the Internet and provides users with highly reliable review summaries. This system has the function of generating emotion scores from reviews using an emotion recognition engine and reflecting these emotion scores in summaries. The detailed configuration and operation of the system are described below.

[1618] System configuration

[1619] The system mainly consists of the following components:

[1620] 1. Server:

[1621] How to collect review information: The server launches a web crawler to automatically collect review information from specific websites. This crawls through specified web pages and extracts data such as the review text, rating score, reviewer information, and posting date and time.

[1622] Analysis method: We use natural language processing (NLP) technology to analyze the collected review information, specifically tokenization, part-of-speech tagging, dependency analysis, and keyword extraction.

[1623] Emotion Recognition Engine: Recognizes emotions from the review text and generates an emotion score. The emotion engine determines the user's emotion (e.g., joy, anger, sadness, etc.) based on the context and keywords in the review text.

[1624] 2. Terminal:

[1625] Summary generation method: The server analyzes the data and generates a sentiment score to generate a reliable review summary, which includes the distribution of rating scores, representative examples of positive and negative opinions, overall rating trends, and sentiment scores.

[1626] User interface: The user can view and refer to the review summary received from the server through the terminal.

[1627] Software and hardware used in the implementation

[1628] Hardware: Smartphones, tablets, PCs

[1629] software:

[1630] Python: a programming language

[1631] BeautifulSoup: A library for web crawling

[1632] requests: HTTP request library

[1633] nltk: A natural language processing library

[1634] SentimentIntensityAnalyzer: An nltk module for sentiment analysis.

[1635] Specific examples

[1636] If a user wants to check reviews of the latest smartphones, they first search for "latest smartphones" within the app. This search request is sent to a server, which collects review information from the specified online shopping site. The collected reviews are analyzed using natural language processing technology to extract important keywords and evaluation scores. An emotion recognition engine is also used to generate an emotion score from the review text. Finally, a reliable review summary is generated based on the analysis results and emotion score, and is displayed through the device's user interface.

[1637] Prompt Sentence Examples

[1638] "Collect reviews of this smartphone and display an overall rating summary based on the analysis results and sentiment score."

[1639] In this way, users can understand not only the overall evaluation of the review information but also the emotional trends contained in each review, enabling them to make highly reliable purchasing decisions.

[1640] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1641] Step 1:

[1642] A user sends a request to the server from their device to collect reviews about a specific product. The input is the user's request, including the product name and category. The server receives this request.

[1643] Step 2:

[1644] The server launches a web crawler and crawls designated websites to collect review information. The input is the user's request, and the output is the collected review information. Specifically, the server uses the crawler to analyze web pages and extract the review text, rating score, reviewer information, and posting date and time.

[1645] Step 3:

[1646] The server stores the collected review information in a database. The input is the collected review information, and the output is analyzable data stored in the database. The server stores the review information by dividing it into fields.

[1647] Step 4:

[1648] The server applies natural language processing (NLP) techniques to the stored review information. The input is the review information read from the database, and the output is the parsed review content. Specifically, the server tokenizes each review text and performs part-of-speech tagging, dependency analysis, keyword extraction, etc.

[1649] Step 5:

[1650] The server categorizes reviews based on the results of natural language processing and categorizes positive and negative reviews based on their rating scores. The input is the analyzed review content, and the output is the categorized reviews. The server also calculates the distribution of rating scores.

[1651] Step 6:

[1652] The server uses an emotion recognition engine to analyze the review text and generate an emotion score. The input is the categorized review content, and the output is an emotion score for each review. Specifically, the server calculates an emotion score (e.g., joy, anger, sadness, etc.) based on the context and keywords in the review text.

[1653] Step 7:

[1654] The server generates a review summary based on the analysis results and sentiment scores. The input is the review analysis results and sentiment scores, and the output is a review summary. The server creates a comprehensive summary that includes the distribution of rating scores, representative examples of positive and negative opinions, overall rating trends, and sentiment scores.

[1655] Step 8:

[1656] When a user requests summary information, the server generates a review summary and sends it to the terminal. The input is the user's summary request, and the output is the summary information.

[1657] Step 9:

[1658] The terminal displays the review summaries received from the server on the user interface. The input is the summary information sent from the server, and the output is the review summary displayed on the user interface. This allows the user to refer to the product ratings and make a purchasing decision.

[1659] 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.

[1660] 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.

[1661] 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.

[1662] 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.

[1663] 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.

[1664] 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.

[1665] 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).

[1666] 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.

[1667] 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."

[1668] 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.

[1669] 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).

[1670] 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.

[1671] 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.

[1672] 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.

[1673] 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.

[1674] 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.

[1675] 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.

[1676] 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.

[1677] 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.

[1678] 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.

[1679] 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.

[1680] The following is further disclosed regarding the above embodiment.

[1681] (Claim 1)

[1682] A means for automatically collecting review information on the Internet;

[1683] A method for analyzing collected review information and extracting key keywords and evaluation scores,

[1684] means for generating a summary of the review based on the analyzed information and displaying it through a user interface;

[1685] A system including:

[1686] (Claim 2)

[1687] 2. The system according to claim 1, wherein the means for collecting the review information is a crawler that patrols specific websites.

[1688] (Claim 3)

[1689] 2. The system according to claim 1, wherein the analysis means analyzes the review content using natural language processing technology.

[1690] "Example 1"

[1691] (Claim 1)

[1692] means for receiving a request from a user to collect review information about a particular product or service;

[1693] A means for automatically collecting review information on the Internet;

[1694] The collected review information is analyzed, and the review text is tokenized using natural language processing technology, followed by part-of-speech tagging and dependency analysis to extract key keywords and evaluation scores.

[1695] a means for generating a summary of the reviews based on the analyzed information and displaying the summary including the distribution of rating scores, representative positive and negative opinions, and overall rating trends through a user interface;

[1696] means for displaying summary information on a user interface in response to a user request;

[1697] A system including:

[1698] (Claim 2)

[1699] 2. The system according to claim 1, wherein the means for collecting the review information is a crawler that patrols specific websites.

[1700] (Claim 3)

[1701] 2. The system according to claim 1, wherein the analysis means uses natural language processing technology to tokenize the review content, tag parts of speech, and perform dependency analysis.

[1702] "Application Example 1"

[1703] (Claim 1)

[1704] A means for automatically collecting evaluation information on the Internet;

[1705] A means of analyzing the collected evaluation information and extracting key keywords and evaluation scores;

[1706] means for generating an assessment summary based on the analyzed information and displaying it through a user interface;

[1707] an input means for a user to search for evaluation information on a specific product;

[1708] a means for generating a summary based on the evaluation information and displaying the summary in a visually easy-to-understand format;

[1709] ...

[1710] A system including:

[1711] (Claim 2)

[1712] 2. The system according to claim 1, wherein the means for collecting the evaluation information is a web crawler that crawls specific websites.

[1713] (Claim 3)

[1714] The system according to claim 1, wherein the analysis means analyzes the evaluation content using natural language processing technology, separates the evaluation information into positive and negative opinions, and generates a distribution of evaluation scores and an overall evaluation trend.

[1715] "Example 2: Combining Emotion Engines"

[1716] (Claim 1)

[1717] A means for automatically collecting review information on the Internet;

[1718] A method for analyzing collected review information and extracting key keywords and evaluation scores,

[1719] means for recognizing a user's emotion based on the analyzed information and generating an emotion score;

[1720] means for generating and displaying a summary of the reviews through a user interface;

[1721] A system including:

[1722] (Claim 2)

[1723] The system according to claim 1, wherein review information is collected using a crawler that patrols specific websites.

[1724] (Claim 3)

[1725] The system of claim 1, wherein the review content is analyzed using natural language processing technology.

[1726] "Application example 2 when combining emotion engines"

[1727] (Claim 1)

[1728] A means for automatically collecting review information on the Internet;

[1729] A method for analyzing collected review information and extracting key keywords and evaluation scores,

[1730] means for generating a summary of the review based on the analyzed information and displaying it through a user interface;

[1731] an emotion recognition engine that generates an emotion score from the review text, and a means for reflecting the generated emotion score in the summary;

[1732] A system including:

[1733] (Claim 2)

[1734] 2. The system according to claim 1, wherein the means for collecting the review information is a crawler that patrols specific websites.

[1735] (Claim 3)

[1736] 2. The system according to claim 1, wherein the analysis means analyzes the review content using natural language processing technology and performs sentiment analysis of the review information using a generative AI model. [Explanation of symbols]

[1737] 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 automatically collecting review information on the Internet; A method for analyzing collected review information and extracting key keywords and evaluation scores, means for generating a summary of the review based on the analyzed information and displaying it through a user interface; A system including:

2. 2. The system according to claim 1, wherein the means for collecting the review information is a crawler that patrols specific websites.

3. 2. The system according to claim 1, wherein the analysis means analyzes the review content using natural language processing technology.

Images