System

Abstract

A system is provided.SOLUTION: A system comprising: a data collection means; a natural language processing means; an inappropriate review exclusion means; a user profile analysis means; and a review providing means.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] Conventional online review platforms are plagued by spam reviews, bot posts, inappropriate reviews, and misleading reviews, making it difficult for consumers to obtain accurate information and increasing the likelihood of businesses receiving unfair reviews. This situation can negatively impact consumer decision-making and cause significant damage to the reputations of businesses and stores. This invention aims to solve these problems and provide only reliable review information. [Means for solving the problem]

[0005] To solve the above problems, the present invention provides a system including the following means: a data collection means for collecting review information from an online review platform; a natural language processing means for analyzing the collected reviews to determine spam characteristics, grammatical consistency, and sentiment analysis; an inappropriate review exclusion means for filtering out spam-like or unnatural reviews; a user profile analysis means for analyzing reviewer profiles to detect unnatural activities; and finally, a review providing means for providing only reliable reviews. This configuration allows consumers to make decisions based on accurate information and allows companies and stores to receive fair evaluations.

[0006] "Data Collection Means" means the means for obtaining and collecting review information from online review platforms.

[0007] "Natural language processing means" refers to means for analyzing collected review information and performing spam characteristics, grammatical consistency, sentiment analysis, etc.

[0008] "Methods for filtering out inappropriate reviews" are measures to identify spam and unnatural reviews based on the analysis results and filter them out.

[0009] "User profile analysis means" refers to means for analyzing reviewer profile information and detecting unusual behavior or activity.

[0010] The "review providing means" is a means for providing filtered and reliable review information to users. [Brief explanation of the drawings]

[0011] [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

[0012] 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.

[0013] First, the terms used in the following description will be explained.

[0014] 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).

[0015] 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.

[0016] 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.

[0017] 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.

[0018] 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."

[0019] [First embodiment]

[0020] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0021] 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.

[0022] 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).

[0023] 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.

[0024] 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.

[0025] 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.

[0026] 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.

[0027] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0028] 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.

[0029] 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.

[0030] 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.

[0031] 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."

[0032] This invention is a system for reliably filtering review information collected from an online review platform. The system operates as follows between a server, a terminal, and a user.

[0033] First, the server collects review information from online review platforms. For example, it uses APIs from Google Maps or other review platforms to obtain reviews about specified stores or services. The collected review information is then stored as a dataset.

[0034] The server then analyzes the collected reviews using natural language processing technology. Specifically, it uses a natural language processing library to analyze the text of each review and perform spam characteristics, grammatical consistency, sentiment analysis, etc. This analysis provides information such as whether each review is spam or contains unnatural content.

[0035] Next, the server filters out inappropriate reviews. This filter removes spam-like reviews, unnatural grammar, or extreme emotional expressions from the list. For example, it filters out reviews that contain the same text repeatedly or that are posted in large numbers in a short period of time.

[0036] The server then analyzes the reviewer's user profile to detect whether the reviewer is engaging in unusual activity, such as multiple reviews submitted by the same user in a short period of time, or frequent reviews submitted in a particular location. Reviews submitted by reviewers with such unusual activity are also filtered out.

[0037] Finally, the server provides the filtered, reliable reviews to the user. When a user requests a review of a specific store or service from their device, the server returns the filtered reviews. This allows users to make decisions based on reliable reviews, and companies and stores can also receive fair evaluations.

[0038] Examples:

[0039] A user uses a device to search for reviews of "restaurants in Tokyo."

[0040] The server collects reviews from Google Maps and other review platforms.

[0041] The collected reviews are analyzed using natural language processing technology to detect spam characteristics and unnatural content.

[0042] The server filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual activity.

[0043] Only trusted reviews resulting from the filtering are provided to the user.

[0044] This is how the system can be implemented. This process allows consumers to make informed decisions and businesses to receive fair reviews.

[0045] The processing flow will be explained below.

[0046] Step 1:

[0047] The server collects review information from online review platforms. Specifically, the server sends a request to the API of Google Maps or other review platforms to retrieve reviews of the specified store or service. The review information returned from the API includes details such as the user ID, review content, rating, and posting date and time.

[0048] Step 2:

[0049] The reviews collected by the server are analyzed using natural language processing. Specifically, a natural language processing library (e.g., spaCy or NLTK) is used to analyze the review text and perform spam characteristics, grammatical consistency, and sentiment analysis. Spam characteristics include analyzing repeated phrases, excessive use of links, and non-natural language. Grammar consistency detects grammatical errors and inappropriate phrases, and sentiment analysis evaluates the review's positive, negative, or neutral sentiment.

[0050] Step 3:

[0051] The server identifies and filters out inappropriate reviews. Using inappropriate review filtering methods, it removes reviews with spam characteristics, reviews with extremely poor grammar, and reviews containing extreme emotional expressions from the list. For example, reviews with excessive emotional expressions like "This store is great!!!" and extremely negative reviews like "It was a terrible experience" are filtered out.

[0052] Step 4:

[0053] The server analyzes the reviewer's user profile. It uses user profile analysis tools to analyze the reviewer's profile. Specifically, it considers the reviewer's posting frequency, the geographic location of the posts, and whether the reviewer's account is newly created. This analysis identifies unusual activity when the same user posts a large number of reviews in a short period of time or frequently posts from multiple different locations.

[0054] Step 5:

[0055] The server provides users with filtered, reliable reviews. When a user requests reviews for a specific store or service from their device, the server returns only reliable reviews that have gone through the filtering process described above. For example, when a user searches for reviews for "restaurants in Tokyo," the server collects, analyzes, and filters reviews from Google Maps and other review platforms and displays them to the user.

[0056] These are the processing steps of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0057] Example 1

[0058] 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."

[0059] Conventional online review platforms contain a large number of spam and inappropriate reviews, making it difficult for users to make decisions based on accurate and reliable information. Furthermore, the lack of mechanisms to detect abnormal reviewer activity has led to issues with impartial evaluations.

[0060] 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.

[0061] In this invention, the server includes a data collection unit, a natural language processing unit, an inappropriate review exclusion unit, a user profile analysis unit, and a review providing unit. This allows for a consistent process of review information collection, analysis, filtering, and provision. Specifically, review information is collected from an online review platform and analyzed using natural language processing technology. Furthermore, inappropriate reviews are identified and excluded using a machine learning algorithm, and abnormal reviewer activity is detected using a clustering algorithm, thereby providing users with reliable review information.

[0062] "Data collection means" refers to a function that automatically obtains review information from online review platforms.

[0063] "Natural language processing means" is a technology that analyzes collected review information and performs tasks such as spam characteristics, grammatical consistency, and sentiment analysis.

[0064] "Means for filtering out inappropriate reviews" is a function that identifies and filters out reviews that contain spam or unnatural language from the collected review information.

[0065] "User profile analysis means" is a function for analyzing reviewer activity patterns and detecting unnatural behavior.

[0066] The "review providing means" is a function that provides filtered, reliable review information in a format that is accessible to users.

[0067] An "online review platform" is a system that allows users to post and view reviews of products and services on the Internet.

[0068] "Machine learning algorithm" is a general term for programs and techniques that automatically learn patterns from data and apply them to new data.

[0069] A "clustering algorithm" is a technique for dividing data into groups, and is a method used in particular to detect anomalous data points.

[0070] "Analysis" is the process of processing collected information for a specific purpose to reveal its characteristics and trends.

[0071] "Filtering" is the act of selecting data based on specific conditions and eliminating unnecessary information.

[0072] The present invention provides a system for reliably filtering review information collected from online review platforms, including a data collection unit, a natural language processing unit, a unit for filtering inappropriate reviews, a unit for analyzing user profiles, and a unit for providing reviews.

[0073] First, the server uses data collection means to collect review information from online review platforms. Specifically, it uses APIs to obtain reviews of designated stores and services from Google Maps and other review platforms. It then secures appropriate access rights using API keys and OAuth authentication, and stores the collected review information in a database.

[0074] The server then analyzes the collected review text using natural language processing (NLP) libraries such as Python's NLTK and spaCy, which determine each review's spam characteristics, grammatical consistency, and sentiment analysis to assess its trustworthiness.

[0075] The server then runs an inappropriate review filtering process, which uses machine learning algorithms (e.g., XGBoost) to automatically identify and filter out spam and unnatural language from the analyzed reviews, ensuring that users only see trustworthy reviews.

[0076] The server then uses user profile analysis tools to analyze the reviewer's activity. Clustering algorithms (e.g., DBSCAN) are used to detect anomalous behavior, such as reviewers posting a large number of reviews in a short period of time or frequently posting reviews in a particular location. Reviews from reviewers with abnormal activity are also rejected.

[0077] Finally, the server provides filtered and reliable reviews to users through a review providing means. When a user requests a review of a specific store or service from their device, the server can return only reliable reviews. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0078] As a concrete example, consider a case where a user searches for reviews of "restaurants in Tokyo." The server collects reviews from Google Maps and other review platforms and analyzes them using natural language processing technology. It then filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual behavior. Finally, the filtered, reliable reviews are provided to the user.

[0079] Example prompt sentence:

[0080] "Show me restaurant reviews in Tokyo."

[0081] This process allows users to make decisions based on accurate and reliable information.

[0082] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0083] Step 1:

[0084] The server starts collecting data. It uses the API of the online review platform to obtain review information about the specified store or service. It receives the store or service ID, API key, and authentication information as input, and generates JSON data of the obtained review information as output. This data is then stored in a database.

[0085] Step 2:

[0086] The server applies natural language processing technology. It uses natural language processing libraries (such as NLTK and spaCy) to analyze the collected review information (text data). It takes the review text data as input and outputs the results of spam characteristics, grammatical consistency, and sentiment analysis. Specifically, it performs sentence tokenization, morphological analysis, and sentiment score calculation.

[0087] Step 3:

[0088] The server filters out inappropriate reviews. It uses machine learning algorithms (such as XGBoost) to classify reviews with spam characteristics or unnatural grammar as inappropriate reviews. It takes the results of natural language processing as input and identifies and filters out reviews with scores that are deemed inappropriate. The output is a filtered list of appropriate reviews.

[0089] Step 4:

[0090] The server analyzes the reviewer's user profile, examines the reviewer's posting frequency and activity patterns, and uses a clustering algorithm (such as DBSCAN) to detect fraudulent activity. It takes the reviewer's posting data (posting date and time, location, etc.) as input and filters out reviews from reviewers who are determined to be anomalous. The output is a list of highly reliable reviews that have been further filtered.

[0091] Step 5:

[0092] The server provides the filtered reviews to the user. When a user requests reviews of a specific store or service from their device, the server returns highly reliable reviews. The user inputs the ID of the store or service in question, and the output is highly reliable review information. This allows users to make decisions based on accurate and reliable information.

[0093] Through the above steps, the system can reliably filter review information collected from online review platforms and provide accurate information to users.

[0094] (Application example 1)

[0095] 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."

[0096] Some reviews provided on online review platforms are inappropriate or unreliable, making it difficult for users to make decisions based on accurate information. This problem is particularly pronounced when purchasing products on e-commerce platforms, where users need to make purchasing decisions based on unbiased and reliable reviews.

[0097] 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.

[0098] In this invention, the server includes a data collection means, a natural language processing means, an inappropriate review exclusion means, a user profile analysis means, a reliable review providing means, a product identification means, and a review data display means, thereby providing reliable reviews to users and enabling users to make decisions based on more accurate information.

[0099] "Data Collection Means" means means for collecting review information from online review platforms or e-commerce platforms.

[0100] The "natural language processing means" is a means for analyzing collected reviews and performing spam characteristics, grammatical consistency, sentiment analysis, and context analysis.

[0101] "Inappropriate review filtering" refers to filtering out reviews that have spam characteristics or contain unnatural grammar or extreme emotional expressions.

[0102] "User profile analysis means" is a means of analyzing whether reviewers are engaging in unnatural activities.

[0103] A "means for providing reliable reviews" is a means for providing filtered, reliable reviews to users.

[0104] A "product identification means" is a means that allows a user to search for reviews about a particular product.

[0105] The "review data display means" is a means for visually displaying reliable reviews to the user.

[0106] This invention is a system that reliably filters review information collected from online review platforms or e-commerce platforms and provides users with the most reliable reviews. Implementing this system requires a server, a user terminal, and a program for linking them.

[0107] 1. System Program Overview

[0108] The server executes a program including a data collection means, a natural language processing means, a means for excluding inappropriate reviews, a means for analyzing user profiles, a means for providing reliable reviews, a means for identifying products, and a means for displaying review data. The program performs the following processes:

[0109] 2. Hardware and Software Used

[0110] Hardware:

[0111] Cloud server: The central hardware that collects and analyzes data.

[0112] Smartphone: The device where users search for reviews and view filtered reviews.

[0113] software:

[0114] Python: A programming language for natural language processing and data analysis.

[0115] TextBlob: A natural language processing library for performing sentiment analysis on reviews.

[0116] NLTK: A natural language processing library used to analyze text data.

[0117] Requests: An HTTP library for collecting review information through an API.

[0118] 3. Program processing explanation

[0119] The server uses the data collection means to collect review information from the online review platform or the e-commerce platform via an API, and the collected review information is stored on the server as a data set.

[0120] Next, we analyze the collected reviews using natural language processing tools. Specifically, we use TextBlob to analyze the sentiment and grammatical consistency of each review, and to detect spam characteristics. We also use NLTK for contextual analysis.

[0121] As a means of excluding inappropriate reviews, we remove reviews that have spam characteristics, reviews that lack grammar consistency, and reviews that contain extreme emotional expressions from the list. We also use user profile analysis to detect unusual reviewer activity and exclude such reviews.

[0122] Finally, the reliable review providing means provides filtered reliable reviews to the user's smartphone. When the user searches for a specific product, the product is identified by the product identification means and visually displayed by the review data display means.

[0123] 4. Examples of concrete examples and prompts

[0124] Example of a user purchasing an iPhone 12:

[0125] A user searches for "iPhone 12" reviews on their smartphone.

[0126] A server collects reviews from an online review platform.

[0127] The collected reviews are analyzed using natural language processing technology to determine spam characteristics, grammatical consistency, sentiment analysis, and context analysis.

[0128] Inappropriate or unreliable reviews will be filtered out.

[0129] Appropriate reviews are filtered and presented to users as trusted reviews.

[0130] Example prompt for a generative AI model:

[0131] Product being reviewed: iPhone 12

[0132] Review List:

[0133] 1. Review ID: 001

[0134] It says: "This product is great. I'm happy with the price and quality."

[0135] Sentiment analysis result: Positive polarity (0.8)

[0136] 2. Review ID: 002

[0137] Content: "spam, fake review"

[0138] Sentiment analysis results: Neutral (0)

[0139] 3. Review ID: 003

[0140] What it said: "Very good product, but a little pricey. Service was good."

[0141] Sentiment analysis result: Positive polarity (0.5)

[0142] This way, users can make decisions based on reliable reviews.

[0143] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0144] Step 1:

[0145] Data collection

[0146] The server uses the data collection means to obtain review information from online review platforms or e-commerce platforms via API. The input is the API endpoint of each platform, and the output is a dataset of collected reviews. Specifically, it sends an HTTP request and stores the obtained data in JSON format on the server.

[0147] Step 2:

[0148] Natural Language Processing

[0149] The server uses natural language processing tools to analyze the collected reviews. The input is a dataset of reviews obtained by the data collection tools, and the output is a dataset of analyzed reviews. Specifically, it uses the TextBlob library to analyze the text of each review and perform sentiment analysis (positive, negative, neutral), grammatical consistency checks, and spam detection.

[0150] Step 3:

[0151] Inappropriate review exclusion

[0152] The server uses inappropriate review filtering methods to filter out reviews with spam characteristics, unnatural grammar, or extreme emotional expressions. The input is a dataset of reviews analyzed using natural language processing methods, and the output is a dataset of reliable reviews. Specifically, it filters out reviews that have been detected as spam characteristics, grammatically incorrect reviews, and reviews with extremely biased emotions.

[0153] Step 4:

[0154] User profile analysis

[0155] The server uses user profile analysis to analyze reviewer activity and detect unusual activity. The input is a filtered review dataset, and the output is a more reliable review dataset. Specifically, it detects cases where a large number of reviews are posted by the same user in a short period of time, or where reviews are frequently posted in a specific location, and excludes those reviews.

[0156] Step 5:

[0157] Providing reliable reviews

[0158] The server uses the reliable review providing means to provide filtered reliable reviews to the user. The input is the final filtered review dataset, and the output is the reliable reviews displayed on the user's device. Specifically, based on the user's request, the server retrieves reliable reviews related to a specific product from the database and sends them in JSON format to the user's smartphone.

[0159] Step 6:

[0160] Product Identification

[0161] The user's device uses a product identification means to search for a specific product. The input is the product name and product ID searched by the user, and the output is a list of reviews related to that product. Specifically, when the user enters the product name into the smartphone app, the ID corresponding to that product is sent to the server.

[0162] Step 7:

[0163] Review data display

[0164] The user's device visually displays the reliable reviews using the review data display means. The input is a dataset of reliable reviews obtained from the server, and the output is review information displayed on the user's smartphone screen. Specifically, the application formats the obtained reviews and displays them in a format that is easy for the user to view.

[0165] The above processing steps allow users to select products with confidence based on highly reliable reviews.

[0166] 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.

[0167] This invention combines a system that reliably filters review information collected from online review platforms with an emotion engine that recognizes user emotions. The system operates as follows between a server, a terminal, and a user.

[0168] First, the server collects review information from online review platforms. Specifically, the server uses a specific API to obtain review data from Google Maps and other review platforms. The collected data includes user IDs, review content, rating points, posting date and time, etc.

[0169] The server then analyzes the collected review information using natural language processing technology. Using natural language processing techniques, the review text undergoes spam filtering, grammar consistency checks, and sentiment analysis. This analysis identifies whether each review is spam or contains unnatural content.

[0170] The server then uses a sentiment engine to analyze the text data in the review and identify positive, negative, or neutral sentiment. The sentiment engine evaluates the review's sentiment by analyzing the frequency of specific keywords and phrases, the positive / negative tendencies of the context, and other factors. During this process, if a review is very positive or very negative, it is identified as spam or unnatural.

[0171] Next, the server identifies and filters out inappropriate reviews. Inappropriate review filtering measures remove reviews with spam characteristics, poor grammar, or extreme sentiment from the list. For example, reviews containing extreme sentiment such as "This restaurant is amazing!!!" or "I'll never go there again."

[0172] The server then analyzes the reviewer's user profile, using user profile analysis tools to analyze the reviewer's posting frequency, geographic location, and the recency of the reviewer account, among other things. This analysis detects unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting in multiple locations.

[0173] Finally, the server provides users with filtered, reliable reviews. When a user requests a review of a specific store or service from their device, the server provides only reliable reviews. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0174] Examples:

[0175] A user uses a device to search for reviews of "restaurants in Tokyo."

[0176] The server collects reviews from Google Maps and other review platforms.

[0177] The collected reviews are analyzed using natural language processing technology and an emotion engine to detect spam characteristics and extreme emotions.

[0178] The server filters out inappropriate reviews and analyzes user profiles to detect unusual activity.

[0179] Only trusted reviews resulting from the filtering are provided to the user.

[0180] This is the implementation form of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0181] The processing flow will be explained below.

[0182] Step 1:

[0183] The server collects review information from online review platforms. Specifically, it uses the APIs of Google Maps and other review platforms to obtain reviews of specified stores and services. This review information includes user IDs, review content, ratings, posting dates, etc.

[0184] Step 2:

[0185] The server analyzes the collected reviews using natural language processing. Specifically, it uses a natural language processing library to analyze the text data and perform spam characteristics, grammatical consistency, and sentiment analysis. For example, it detects the presence of specific spam keywords and grammatical errors, and determines whether the sentiment of the comment is positive, negative, or neutral.

[0186] Step 3:

[0187] The server uses a sentiment engine to evaluate the detailed sentiment of the analyzed reviews. The sentiment engine analyzes specific keywords and phrases within the review to generate an overall sentiment score. Based on this score, the server identifies whether the review is excessively positive or negative. Based on this result, the server can determine whether an overly emotional review is inappropriate.

[0188] Step 4:

[0189] The server identifies and filters out inappropriate reviews. It uses inappropriate review filtering to remove reviews that have spammy characteristics, poor grammar, or extreme sentiments that are identified by the sentiment engine. For example, reviews like "This store is great!!!" and "I'll never go there again" are filtered out.

[0190] Step 5:

[0191] The server analyzes the reviewer's user profile. Using user profile analysis methods, it checks the reviewer's posting frequency, geographic location, account recency, etc. This profile analysis identifies unusual activity when the same user posts a large number of reviews in a short period of time or frequently posts reviews in different locations.

[0192] Step 6:

[0193] The server provides users with filtered, reliable reviews. When a user requests reviews for a specific store or service using their device, the server returns only the filtered, reliable reviews, allowing users to make decisions based on accurate information.

[0194] Examples:

[0195] A user uses a device to search for reviews of "restaurants in Tokyo."

[0196] The server collects review information from Google Maps and other review platforms.

[0197] The collected reviews are analyzed using natural language processing tools and an emotion engine to detect spam characteristics and emotional extremes.

[0198] The server filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual behavior.

[0199] The server provides filtered and trusted reviews to the user.

[0200] These are the specific processing steps of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0201] Example 2

[0202] 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."

[0203] Online review platforms are often filled with reviews containing unreliable information, spam reviews, and unnatural emotional expressions. The presence of such inappropriate reviews makes it difficult for users to make accurate decisions. Furthermore, reviewer credibility is not evaluated, and inappropriate activity may be overlooked. Therefore, a system is needed that provides reliable reviews and enables users to make accurate decisions.

[0204] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0205] In this invention, the server includes a data collection means, a natural language processing means, a sentiment analysis means, an inappropriate review excluding means, a user profile analysis means, and a review providing means, thereby enabling the server to reliably filter review information collected from online review platforms and provide accurate and fair reviews to users.

[0206] "Data collection means" refers to a device that has the function of collecting review information from an online review platform.

[0207] The "natural language processing means" is a device that has the function of analyzing collected reviews and performing spam characteristics, grammatical consistency, and sentiment analysis.

[0208] The "sentiment analysis means" is a device that has the function of analyzing text data in reviews and identifying positive, negative, and neutral emotions.

[0209] An "inappropriate review filtering method" is a device that has the function of excluding reviews that have spam characteristics, poor grammar, or extreme emotions from the list.

[0210] A "user profile analysis means" is a device that has the function of analyzing a reviewer's posting frequency, geographic location, the recency of the reviewer account, etc., and detecting any unusual activity.

[0211] The "review providing means" is a device that has the function of providing filtered and reliable reviews to users.

[0212] This invention combines a system for reliably filtering review information collected from online review platforms with an emotion engine that recognizes user emotions. The system operates among a server, a terminal, and a user as follows: The server processes and calculates data using specific hardware and software.

[0213] First, the server collects review information from an online review platform. Specifically, the server obtains review data using a specific API (e.g., online review platform API). The collected data includes user IDs, review content, rating points, posting date and time, etc.

[0214] The server then analyzes the collected review information using natural language processing technology. The natural language processing uses Python natural language processing libraries such as NLTK and spaCy. The server then performs spam filtering, grammar consistency checks, and sentiment analysis. This allows it to determine whether each review is spam or contains unnatural content.

[0215] Additionally, the server uses a sentiment engine (e.g., natural language sentiment API) to analyze the text data in the review and identify positive, negative, or neutral sentiment. The sentiment engine evaluates the sentiment of the review by analyzing the frequency of specific keywords and phrases, the positive / negative tendencies of the context, etc. During this process, if a review is very positive or very negative, it is identified as spam or unnatural.

[0216] Next, the server identifies and filters out inappropriate reviews. Inappropriate review filtering measures remove reviews with spam characteristics, poor grammar, or extreme sentiment from the list. For example, reviews containing extreme sentiment such as "This restaurant is amazing!!!" or "I'll never go there again."

[0217] The server then analyzes the reviewer's user profile. Using user profile analysis methods, the server analyzes the reviewer's posting frequency, geographic location, and the recency of the reviewer account. For example, it detects unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting in multiple locations. This is done using the database management systems MySQL and Apache Cassandra.

[0218] Finally, the server provides the filtered, trusted reviews to the user. When a user requests a review of a specific store or service from their device, the server sends only trusted reviews to the device. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0219] Examples:

[0220] A user uses a device to search for reviews of "restaurants in Tokyo."

[0221] The server collects reviews from an online review platform API.

[0222] The collected reviews are analyzed using natural language processing technology and an emotion engine to detect spam characteristics and extreme emotions.

[0223] The server filters out inappropriate reviews and analyzes user profiles to detect unusual activity.

[0224] Only trusted reviews resulting from the filtering are provided to the user.

[0225] Example prompt sentence:

[0226] "Write a program to identify users who submit a large number of reviews in a certain period of time and detect any unusual activity."

[0227] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0228] Step 1:

[0229] The server uses the API of the online review platform to collect review information. As input, it receives data such as the user ID, review content, rating, and posting date and time. Based on this, it stores the review information in a database. This is how the review information is collected.

[0230] Step 2:

[0231] The server analyzes the collected review information using a Python natural language processing library (e.g., NLTK or spaCy). The review content is given as input. First, the review text is preprocessed (tokenization, stop word removal), then spam filtering (Bayesian filtering, etc.) and grammatical consistency check (generating a parse tree). This allows the spam characteristics and grammatical consistency to be evaluated.

[0232] Step 3:

[0233] The server analyzes the sentiment of the reviews using a sentiment engine (e.g., a natural language sentiment API). The preprocessed review content is given as input. The sentiment engine analyzes the frequency and context of specific keywords and phrases, and outputs a sentiment score as positive, negative, or neutral. This classifies the sentiment of each review.

[0234] Step 4:

[0235] The server identifies and filters out inappropriate reviews. It receives inputs such as sentiment scores, spam characteristics, and grammatical consistency assessment results. The inappropriate review filtering method filters out reviews with spam characteristics or extreme sentiment, eliminating unreliable reviews.

[0236] Step 5:

[0237] The server analyzes the reviewer's user profile. As input, it takes information such as the reviewer's posting frequency, geographic location, and account recency. It uses a database management system (e.g., MySQL or Apache Cassandra) to detect unusual activity, such as users posting a large number of reviews in a short period of time or users posting reviews frequently in different locations. This allows it to assess the reviewer's trustworthiness.

[0238] Step 6:

[0239] A user requests reviews of a specific store or service using a device. A search keyword (e.g., "restaurants in Tokyo") is given as input. The server then provides filtered, trusted reviews, allowing users to make decisions based on accurate information.

[0240] (Application example 2)

[0241] 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."

[0242] Conventional review systems often contain spam reviews or reviews with extreme emotional expressions, making it difficult for users to obtain reliable reviews. Furthermore, the credibility of reviews based on reviewer profiles and activity histories is often insufficient, negatively impacting users' judgments. Furthermore, there has been no consistent method to effectively filter these reviews and provide only reliable reviews.

[0243] 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 a data collection means, a natural language processing means, an inappropriate review exclusion means, a user profile analysis means, a review providing means, a sentiment analysis means, and an application means to be installed on the terminal. This makes it possible to analyze review information collected from an online review platform, exclude spam and inappropriate reviews, and display only reliable reviews on the user's terminal.

[0244] "Data Collection Instruments" means instruments capable of collecting review information from online review platforms.

[0245] "Natural language processing means" refers to means for analyzing collected review information and performing spam characteristics, grammatical consistency, and sentiment analysis.

[0246] "Measures to filter out inappropriate reviews" are measures to detect and filter out reviews that have spam-like characteristics, reviews that contain extreme emotional expressions, and reviews with unnatural grammar.

[0247] "User profile analysis means" refers to means for analyzing profile information such as reviewer posting frequency, geographic location, and account recency.

[0248] The "review providing means" is a means for providing filtered and reliable review information to users.

[0249] The "sentiment analysis means" is a means for analyzing the content of a review and identifying positive, negative, or neutral sentiment.

[0250] The "application means installed on the terminal" refers to a means including an application that provides only reliable reviews when review information is displayed on the terminal.

[0251] This invention relates to a system that reliably filters review information collected from online review platforms and provides it to users. This system mainly consists of three elements: a server, a terminal, and a user.

[0252] The server collects review information from online review platforms. Specifically, the server uses a specific API to obtain review information from various review platforms. The collected data includes user IDs, review content, rating points, posting date and time, etc.

[0253] The server analyzes the collected review information using natural language processing. Utilizing natural language processing technology, the review content is subjected to spam filtering, grammar consistency checks, and sentiment analysis. This analysis determines whether each review is spam or contains unnatural content.

[0254] The server then uses sentiment analysis to analyze the review content and identify positive, negative, or neutral sentiment. Sentiment analysis is performed by analyzing the frequency of specific keywords and phrases, as well as contextual sentiment trends. During this process, if a review is extremely positive or negative, it is identified as spam or unnatural.

[0255] The server then detects and filters out inappropriate reviews, such as spam, reviews with extreme emotional expressions, and reviews with poor grammar, eliminating reviews with low credibility from the list.

[0256] The server also analyzes reviewer profiles using user profile analysis tools, which analyze reviewer posting frequency, geographic location, account recency, etc., to detect unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting from multiple different locations.

[0257] The terminal displays reviews of products and services searched for by the user through an application means installed on the terminal. The application has a function of displaying only reliable reviews sent from the server, allowing the user to make decisions based on reliable reviews.

[0258] For example, a user searches for a smartphone case on an online shopping site using a device. The server collects review information and analyzes the reviews using sentiment analysis and natural language processing. Inappropriate reviews are filtered out, and only reliable reviews are provided to the user's device after analyzing the user profile.

[0259] The program to realize this process is implemented using Python, Google Cloud Natural Language API, IBM Watson Tone Analyzer API, etc. This system enables users to make accurate decisions based on reliable reviews, and also enables companies and service providers to receive fair evaluations.

[0260] Example prompt: "Create a Python program that uses the data obtained from the review collection API call to analyze the review content using the Google Cloud Natural Language API and IBM Watson Tone Analyzer API, filter out spam and inappropriate reviews, and provide only trustworthy reviews to users."

[0261] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0262] Step 1:

[0263] Collecting review information

[0264] Subject: Server

[0265] Description: The server uses the API of an online review platform to collect review information for a specific product. For example, to obtain reviews for a specific product, the server calls the API and obtains data including the review content, user ID, rating, posting date and time, etc.

[0266] Input: Product ID given to the API

[0267] Output: A list of retrieved review information

[0268] Step 2:

[0269] Review analysis using natural language processing

[0270] Subject: Server

[0271] Description: The server analyzes the collected review information using a natural language processing engine (e.g., Google Cloud Natural Language API). Specifically, it evaluates the credibility of each review by filtering spam, checking grammar consistency, and analyzing sentiment.

[0272] Input: List of review information

[0273] Output: A list of parsed review information (with spam characteristics, grammar consistency, and sentiment scores)

[0274] Step 3:

[0275] Emotion analysis

[0276] Subject: Server

[0277] Description: The server uses a sentiment analysis engine to classify review content as positive, negative, or neutral. This process involves evaluating the frequency of specific keywords and phrases, the positive / negative tendencies of the context, and other factors to generate a sentiment score. Specifically, it uses the IBM Watson Tone Analyzer API.

[0278] Input: A list of parsed review information (with spam characteristics and grammar consistency)

[0279] Output: A list of reviews with sentiment scores

[0280] Step 4:

[0281] Filtering out inappropriate reviews

[0282] Subject: Server

[0283] Description: The server detects and filters out inappropriate reviews based on the sentiment score obtained by the sentiment analysis method. Reviews that are excessively positive or negative, spam-like, or have poor grammar are removed.

[0284] Input: A list of reviews with sentiment scores

[0285] Output: A list of filtered reviews

[0286] Step 5:

[0287] User profile analysis

[0288] Subject: Server

[0289] Description: The server analyzes the reviewer's user profile, which includes posting frequency, geographic location, account recency, etc. If a reviewer posts a large number of reviews in a short period of time or frequently from multiple different locations, this is detected as unusual activity.

[0290] Input: A filtered list of reviews

[0291] Output: A list of reliable reviews

[0292] Step 6:

[0293] Providing a review

[0294] Subject: Terminal

[0295] Description: Provide users with reliable reviews through an application installed on a terminal, which displays reliable review information received from a server to the user.

[0296] Input: A list of reliable reviews

[0297] Output: Reliable review information provided to users

[0298] In this way, the server and the terminal work together to realize a system that provides highly reliable review information to users.

[0299] 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.

[0300] 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.

[0301] 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.

[0302] [Second embodiment]

[0303] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0304] 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.

[0305] 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).

[0306] 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.

[0307] 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.

[0308] 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).

[0309] 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.

[0310] 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.

[0311] 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.

[0312] 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.

[0313] 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.

[0314] 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."

[0315] This invention is a system for reliably filtering review information collected from an online review platform. The system operates as follows between a server, a terminal, and a user.

[0316] First, the server collects review information from online review platforms. For example, it uses APIs from Google Maps or other review platforms to obtain reviews about specified stores or services. The collected review information is then stored as a dataset.

[0317] The server then analyzes the collected reviews using natural language processing technology. Specifically, it uses a natural language processing library to analyze the text of each review and perform spam characteristics, grammatical consistency, sentiment analysis, etc. This analysis provides information such as whether each review is spam or contains unnatural content.

[0318] Next, the server filters out inappropriate reviews. This filter removes spam-like reviews, unnatural grammar, or extreme emotional expressions from the list. For example, it filters out reviews that contain the same text repeatedly or that are posted in large numbers in a short period of time.

[0319] The server then analyzes the reviewer's user profile to detect whether the reviewer is engaging in unusual activity, such as multiple reviews submitted by the same user in a short period of time, or frequent reviews submitted in a particular location. Reviews submitted by reviewers with such unusual activity are also filtered out.

[0320] Finally, the server provides the filtered, reliable reviews to the user. When a user requests a review of a specific store or service from their device, the server returns the filtered reviews. This allows users to make decisions based on reliable reviews, and companies and stores can also receive fair evaluations.

[0321] Examples:

[0322] A user uses a device to search for reviews of "restaurants in Tokyo."

[0323] The server collects reviews from Google Maps and other review platforms.

[0324] The collected reviews are analyzed using natural language processing technology to detect spam characteristics and unnatural content.

[0325] The server filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual activity.

[0326] Only trusted reviews resulting from the filtering are provided to the user.

[0327] This is how the system can be implemented. This process allows consumers to make informed decisions and businesses to receive fair reviews.

[0328] The processing flow will be explained below.

[0329] Step 1:

[0330] The server collects review information from online review platforms. Specifically, the server sends a request to the API of Google Maps or other review platforms to retrieve reviews of the specified store or service. The review information returned from the API includes details such as the user ID, review content, rating, and posting date and time.

[0331] Step 2:

[0332] The reviews collected by the server are analyzed using natural language processing. Specifically, a natural language processing library (e.g., spaCy or NLTK) is used to analyze the review text and perform spam characteristics, grammatical consistency, and sentiment analysis. Spam characteristics include analyzing repeated phrases, excessive use of links, and non-natural language. Grammar consistency detects grammatical errors and inappropriate phrases, and sentiment analysis evaluates the review's positive, negative, or neutral sentiment.

[0333] Step 3:

[0334] The server identifies and filters out inappropriate reviews. Using inappropriate review filtering methods, it removes reviews with spam characteristics, reviews with extremely poor grammar, and reviews containing extreme emotional expressions from the list. For example, reviews with excessive emotional expressions like "This store is great!!!" and extremely negative reviews like "It was a terrible experience" are filtered out.

[0335] Step 4:

[0336] The server analyzes the reviewer's user profile. It uses user profile analysis tools to analyze the reviewer's profile. Specifically, it considers the reviewer's posting frequency, the geographic location of the posts, and whether the reviewer's account is newly created. This analysis identifies unusual activity when the same user posts a large number of reviews in a short period of time or frequently posts from multiple different locations.

[0337] Step 5:

[0338] The server provides users with filtered, reliable reviews. When a user requests reviews for a specific store or service from their device, the server returns only reliable reviews that have gone through the filtering process described above. For example, when a user searches for reviews for "restaurants in Tokyo," the server collects, analyzes, and filters reviews from Google Maps and other review platforms and displays them to the user.

[0339] These are the processing steps of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0340] Example 1

[0341] 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."

[0342] Conventional online review platforms contain a large number of spam and inappropriate reviews, making it difficult for users to make decisions based on accurate and reliable information. Furthermore, the lack of mechanisms to detect abnormal reviewer activity has led to issues with impartial evaluations.

[0343] 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.

[0344] In this invention, the server includes a data collection unit, a natural language processing unit, an inappropriate review exclusion unit, a user profile analysis unit, and a review providing unit. This allows for a consistent process of review information collection, analysis, filtering, and provision. Specifically, review information is collected from an online review platform and analyzed using natural language processing technology. Furthermore, inappropriate reviews are identified and excluded using a machine learning algorithm, and abnormal reviewer activity is detected using a clustering algorithm, thereby providing users with reliable review information.

[0345] "Data collection means" refers to a function that automatically obtains review information from online review platforms.

[0346] "Natural language processing means" is a technology that analyzes collected review information and performs tasks such as spam characteristics, grammatical consistency, and sentiment analysis.

[0347] "Means for filtering out inappropriate reviews" is a function that identifies and filters out reviews that contain spam or unnatural language from the collected review information.

[0348] "User profile analysis means" is a function for analyzing reviewer activity patterns and detecting unnatural behavior.

[0349] The "review providing means" is a function that provides filtered, reliable review information in a format that is accessible to users.

[0350] An "online review platform" is a system that allows users to post and view reviews of products and services on the Internet.

[0351] "Machine learning algorithm" is a general term for programs and techniques that automatically learn patterns from data and apply them to new data.

[0352] A "clustering algorithm" is a technique for dividing data into groups, and is a method used in particular to detect anomalous data points.

[0353] "Analysis" is the process of processing collected information for a specific purpose to reveal its characteristics and trends.

[0354] "Filtering" is the act of selecting data based on specific conditions and eliminating unnecessary information.

[0355] The present invention provides a system for reliably filtering review information collected from online review platforms, including a data collection unit, a natural language processing unit, a unit for filtering inappropriate reviews, a unit for analyzing user profiles, and a unit for providing reviews.

[0356] First, the server uses data collection means to collect review information from online review platforms. Specifically, it uses APIs to obtain reviews of designated stores and services from Google Maps and other review platforms. It then secures appropriate access rights using API keys and OAuth authentication, and stores the collected review information in a database.

[0357] The server then analyzes the collected review text using natural language processing (NLP) libraries such as Python's NLTK and spaCy, which determine each review's spam characteristics, grammatical consistency, and sentiment analysis to assess its trustworthiness.

[0358] The server then runs an inappropriate review filtering process, which uses machine learning algorithms (e.g., XGBoost) to automatically identify and filter out spam and unnatural language from the analyzed reviews, ensuring that users only see trustworthy reviews.

[0359] The server then uses user profile analysis tools to analyze the reviewer's activity. Clustering algorithms (e.g., DBSCAN) are used to detect anomalous behavior, such as reviewers posting a large number of reviews in a short period of time or frequently posting reviews in a particular location. Reviews from reviewers with abnormal activity are also rejected.

[0360] Finally, the server provides filtered and reliable reviews to users through a review providing means. When a user requests a review of a specific store or service from their device, the server can return only reliable reviews. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0361] As a concrete example, consider a case where a user searches for reviews of "restaurants in Tokyo." The server collects reviews from Google Maps and other review platforms and analyzes them using natural language processing technology. It then filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual behavior. Finally, the filtered, reliable reviews are provided to the user.

[0362] Example prompt sentence:

[0363] "Show me restaurant reviews in Tokyo."

[0364] This process allows users to make decisions based on accurate and reliable information.

[0365] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0366] Step 1:

[0367] The server starts collecting data. It uses the API of the online review platform to obtain review information about the specified store or service. It receives the store or service ID, API key, and authentication information as input, and generates JSON data of the obtained review information as output. This data is then stored in a database.

[0368] Step 2:

[0369] The server applies natural language processing technology. It uses natural language processing libraries (such as NLTK and spaCy) to analyze the collected review information (text data). It takes the review text data as input and outputs the results of spam characteristics, grammatical consistency, and sentiment analysis. Specifically, it performs sentence tokenization, morphological analysis, and sentiment score calculation.

[0370] Step 3:

[0371] The server filters out inappropriate reviews. It uses machine learning algorithms (such as XGBoost) to classify reviews with spam characteristics or unnatural grammar as inappropriate reviews. It takes the results of natural language processing as input and identifies and filters out reviews with scores that are deemed inappropriate. The output is a filtered list of appropriate reviews.

[0372] Step 4:

[0373] The server analyzes the reviewer's user profile, examines the reviewer's posting frequency and activity patterns, and uses a clustering algorithm (such as DBSCAN) to detect fraudulent activity. It takes the reviewer's posting data (posting date and time, location, etc.) as input and filters out reviews from reviewers who are determined to be anomalous. The output is a list of highly reliable reviews that have been further filtered.

[0374] Step 5:

[0375] The server provides the filtered reviews to the user. When a user requests reviews of a specific store or service from their device, the server returns highly reliable reviews. The user inputs the ID of the store or service in question, and the output is highly reliable review information. This allows users to make decisions based on accurate and reliable information.

[0376] Through the above steps, the system can reliably filter review information collected from online review platforms and provide accurate information to users.

[0377] (Application example 1)

[0378] 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."

[0379] Some reviews provided on online review platforms are inappropriate or unreliable, making it difficult for users to make decisions based on accurate information. This problem is particularly pronounced when purchasing products on e-commerce platforms, where users need to make purchasing decisions based on unbiased and reliable reviews.

[0380] 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.

[0381] In this invention, the server includes a data collection means, a natural language processing means, an inappropriate review exclusion means, a user profile analysis means, a reliable review providing means, a product identification means, and a review data display means, thereby providing reliable reviews to users and enabling users to make decisions based on more accurate information.

[0382] "Data Collection Means" means means for collecting review information from online review platforms or e-commerce platforms.

[0383] The "natural language processing means" is a means for analyzing collected reviews and performing spam characteristics, grammatical consistency, sentiment analysis, and context analysis.

[0384] "Inappropriate review filtering" refers to filtering out reviews that have spam characteristics or contain unnatural grammar or extreme emotional expressions.

[0385] "User profile analysis means" is a means of analyzing whether reviewers are engaging in unnatural activities.

[0386] A "means for providing reliable reviews" is a means for providing filtered, reliable reviews to users.

[0387] A "product identification means" is a means that allows a user to search for reviews about a particular product.

[0388] The "review data display means" is a means for visually displaying reliable reviews to the user.

[0389] This invention is a system that reliably filters review information collected from online review platforms or e-commerce platforms and provides users with the most reliable reviews. Implementing this system requires a server, a user terminal, and a program for linking them.

[0390] 1. System Program Overview

[0391] The server executes a program including a data collection means, a natural language processing means, a means for excluding inappropriate reviews, a means for analyzing user profiles, a means for providing reliable reviews, a means for identifying products, and a means for displaying review data. The program performs the following processes:

[0392] 2. Hardware and Software Used

[0393] Hardware:

[0394] Cloud server: The central hardware that collects and analyzes data.

[0395] Smartphone: The device where users search for reviews and view filtered reviews.

[0396] software:

[0397] Python: A programming language for natural language processing and data analysis.

[0398] TextBlob: A natural language processing library for performing sentiment analysis on reviews.

[0399] NLTK: A natural language processing library used to analyze text data.

[0400] Requests: An HTTP library for collecting review information through an API.

[0401] 3. Program processing explanation

[0402] The server uses the data collection means to collect review information from the online review platform or the e-commerce platform via an API, and the collected review information is stored on the server as a data set.

[0403] Next, we analyze the collected reviews using natural language processing tools. Specifically, we use TextBlob to analyze the sentiment and grammatical consistency of each review, and to detect spam characteristics. We also use NLTK for contextual analysis.

[0404] As a means of excluding inappropriate reviews, we remove reviews that have spam characteristics, reviews that lack grammar consistency, and reviews that contain extreme emotional expressions from the list. We also use user profile analysis to detect unusual reviewer activity and exclude such reviews.

[0405] Finally, the reliable review providing means provides filtered reliable reviews to the user's smartphone. When the user searches for a specific product, the product is identified by the product identification means and visually displayed by the review data display means.

[0406] 4. Examples of concrete examples and prompts

[0407] Example of a user purchasing an iPhone 12:

[0408] A user searches for "iPhone 12" reviews on their smartphone.

[0409] A server collects reviews from an online review platform.

[0410] The collected reviews are analyzed using natural language processing technology to determine spam characteristics, grammatical consistency, sentiment analysis, and context analysis.

[0411] Inappropriate or unreliable reviews will be filtered out.

[0412] Appropriate reviews are filtered and presented to users as trusted reviews.

[0413] Example prompt for a generative AI model:

[0414] Product being reviewed: iPhone 12

[0415] Review List:

[0416] 1. Review ID: 001

[0417] It says: "This product is great. I'm happy with the price and quality."

[0418] Sentiment analysis result: Positive polarity (0.8)

[0419] 2. Review ID: 002

[0420] Content: "spam, fake review"

[0421] Sentiment analysis results: Neutral (0)

[0422] 3. Review ID: 003

[0423] What it said: "Very good product, but a little pricey. Service was good."

[0424] Sentiment analysis result: Positive polarity (0.5)

[0425] This way, users can make decisions based on reliable reviews.

[0426] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0427] Step 1:

[0428] Data collection

[0429] The server uses the data collection means to obtain review information from online review platforms or e-commerce platforms via API. The input is the API endpoint of each platform, and the output is a dataset of collected reviews. Specifically, it sends an HTTP request and stores the obtained data in JSON format on the server.

[0430] Step 2:

[0431] Natural Language Processing

[0432] The server uses natural language processing tools to analyze the collected reviews. The input is a dataset of reviews obtained by the data collection tools, and the output is a dataset of analyzed reviews. Specifically, it uses the TextBlob library to analyze the text of each review and perform sentiment analysis (positive, negative, neutral), grammatical consistency checks, and spam detection.

[0433] Step 3:

[0434] Inappropriate review exclusion

[0435] The server uses inappropriate review filtering methods to filter out reviews with spam characteristics, unnatural grammar, or extreme emotional expressions. The input is a dataset of reviews analyzed using natural language processing methods, and the output is a dataset of reliable reviews. Specifically, it filters out reviews that have been detected as spam characteristics, grammatically incorrect reviews, and reviews with extremely biased emotions.

[0436] Step 4:

[0437] User profile analysis

[0438] The server uses user profile analysis to analyze reviewer activity and detect unusual activity. The input is a filtered review dataset, and the output is a more reliable review dataset. Specifically, it detects cases where a large number of reviews are posted by the same user in a short period of time, or where reviews are frequently posted in a specific location, and excludes those reviews.

[0439] Step 5:

[0440] Providing reliable reviews

[0441] The server uses the reliable review providing means to provide filtered reliable reviews to the user. The input is the final filtered review dataset, and the output is the reliable reviews displayed on the user's device. Specifically, based on the user's request, the server retrieves reliable reviews related to a specific product from the database and sends them in JSON format to the user's smartphone.

[0442] Step 6:

[0443] Product Identification

[0444] The user's device uses a product identification means to search for a specific product. The input is the product name and product ID searched by the user, and the output is a list of reviews related to that product. Specifically, when the user enters the product name into the smartphone app, the ID corresponding to that product is sent to the server.

[0445] Step 7:

[0446] Review data display

[0447] The user's device visually displays the reliable reviews using the review data display means. The input is a dataset of reliable reviews obtained from the server, and the output is review information displayed on the user's smartphone screen. Specifically, the application formats the obtained reviews and displays them in a format that is easy for the user to view.

[0448] The above processing steps allow users to select products with confidence based on highly reliable reviews.

[0449] 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.

[0450] This invention combines a system that reliably filters review information collected from online review platforms with an emotion engine that recognizes user emotions. The system operates as follows between a server, a terminal, and a user.

[0451] First, the server collects review information from online review platforms. Specifically, the server uses a specific API to obtain review data from Google Maps and other review platforms. The collected data includes user IDs, review content, rating points, posting date and time, etc.

[0452] The server then analyzes the collected review information using natural language processing technology. Using natural language processing techniques, the review text undergoes spam filtering, grammar consistency checks, and sentiment analysis. This analysis identifies whether each review is spam or contains unnatural content.

[0453] The server then uses a sentiment engine to analyze the text data in the review and identify positive, negative, or neutral sentiment. The sentiment engine evaluates the review's sentiment by analyzing the frequency of specific keywords and phrases, the positive / negative tendencies of the context, and other factors. During this process, if a review is very positive or very negative, it is identified as spam or unnatural.

[0454] Next, the server identifies and filters out inappropriate reviews. Inappropriate review filtering measures remove reviews with spam characteristics, poor grammar, or extreme sentiment from the list. For example, reviews containing extreme sentiment such as "This restaurant is amazing!!!" or "I'll never go there again."

[0455] The server then analyzes the reviewer's user profile, using user profile analysis tools to analyze the reviewer's posting frequency, geographic location, and the recency of the reviewer account, among other things. This analysis detects unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting in multiple locations.

[0456] Finally, the server provides users with filtered, reliable reviews. When a user requests a review of a specific store or service from their device, the server provides only reliable reviews. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0457] Examples:

[0458] A user uses a device to search for reviews of "restaurants in Tokyo."

[0459] The server collects reviews from Google Maps and other review platforms.

[0460] The collected reviews are analyzed using natural language processing technology and an emotion engine to detect spam characteristics and extreme emotions.

[0461] The server filters out inappropriate reviews and analyzes user profiles to detect unusual activity.

[0462] Only trusted reviews resulting from the filtering are provided to the user.

[0463] This is the implementation form of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0464] The processing flow will be explained below.

[0465] Step 1:

[0466] The server collects review information from online review platforms. Specifically, it uses the APIs of Google Maps and other review platforms to obtain reviews of specified stores and services. This review information includes user IDs, review content, ratings, posting dates, etc.

[0467] Step 2:

[0468] The server analyzes the collected reviews using natural language processing. Specifically, it uses a natural language processing library to analyze the text data and perform spam characteristics, grammatical consistency, and sentiment analysis. For example, it detects the presence of specific spam keywords and grammatical errors, and determines whether the sentiment of the comment is positive, negative, or neutral.

[0469] Step 3:

[0470] The server uses a sentiment engine to evaluate the detailed sentiment of the analyzed reviews. The sentiment engine analyzes specific keywords and phrases within the review to generate an overall sentiment score. Based on this score, the server identifies whether the review is excessively positive or negative. Based on this result, the server can determine whether an overly emotional review is inappropriate.

[0471] Step 4:

[0472] The server identifies and filters out inappropriate reviews. It uses inappropriate review filtering to remove reviews that have spammy characteristics, poor grammar, or extreme sentiments that are identified by the sentiment engine. For example, reviews like "This store is great!!!" and "I'll never go there again" are filtered out.

[0473] Step 5:

[0474] The server analyzes the reviewer's user profile. Using user profile analysis methods, it checks the reviewer's posting frequency, geographic location, account recency, etc. This profile analysis identifies unusual activity when the same user posts a large number of reviews in a short period of time or frequently posts reviews in different locations.

[0475] Step 6:

[0476] The server provides users with filtered, reliable reviews. When a user requests reviews for a specific store or service using their device, the server returns only the filtered, reliable reviews, allowing users to make decisions based on accurate information.

[0477] Examples:

[0478] A user uses a device to search for reviews of "restaurants in Tokyo."

[0479] The server collects review information from Google Maps and other review platforms.

[0480] The collected reviews are analyzed using natural language processing tools and an emotion engine to detect spam characteristics and emotional extremes.

[0481] The server filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual behavior.

[0482] The server provides filtered and trusted reviews to the user.

[0483] These are the specific processing steps of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0484] Example 2

[0485] 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."

[0486] Online review platforms are often filled with reviews containing unreliable information, spam reviews, and unnatural emotional expressions. The presence of such inappropriate reviews makes it difficult for users to make accurate decisions. Furthermore, reviewer credibility is not evaluated, and inappropriate activity may be overlooked. Therefore, a system is needed that provides reliable reviews and enables users to make accurate decisions.

[0487] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0488] In this invention, the server includes a data collection means, a natural language processing means, a sentiment analysis means, an inappropriate review excluding means, a user profile analysis means, and a review providing means, thereby enabling the server to reliably filter review information collected from online review platforms and provide accurate and fair reviews to users.

[0489] "Data collection means" refers to a device that has the function of collecting review information from an online review platform.

[0490] The "natural language processing means" is a device that has the function of analyzing collected reviews and performing spam characteristics, grammatical consistency, and sentiment analysis.

[0491] The "sentiment analysis means" is a device that has the function of analyzing text data in reviews and identifying positive, negative, and neutral emotions.

[0492] An "inappropriate review filtering method" is a device that has the function of excluding reviews that have spam characteristics, poor grammar, or extreme emotions from the list.

[0493] A "user profile analysis means" is a device that has the function of analyzing a reviewer's posting frequency, geographic location, the recency of the reviewer account, etc., and detecting any unusual activity.

[0494] The "review providing means" is a device that has the function of providing filtered and reliable reviews to users.

[0495] This invention combines a system for reliably filtering review information collected from online review platforms with an emotion engine that recognizes user emotions. The system operates among a server, a terminal, and a user as follows: The server processes and calculates data using specific hardware and software.

[0496] First, the server collects review information from an online review platform. Specifically, the server obtains review data using a specific API (e.g., online review platform API). The collected data includes user IDs, review content, rating points, posting date and time, etc.

[0497] The server then analyzes the collected review information using natural language processing technology. The natural language processing uses Python natural language processing libraries such as NLTK and spaCy. The server then performs spam filtering, grammar consistency checks, and sentiment analysis. This allows it to determine whether each review is spam or contains unnatural content.

[0498] Additionally, the server uses a sentiment engine (e.g., natural language sentiment API) to analyze the text data in the review and identify positive, negative, or neutral sentiment. The sentiment engine evaluates the sentiment of the review by analyzing the frequency of specific keywords and phrases, the positive / negative tendencies of the context, etc. During this process, if a review is very positive or very negative, it is identified as spam or unnatural.

[0499] Next, the server identifies and filters out inappropriate reviews. Inappropriate review filtering measures remove reviews with spam characteristics, poor grammar, or extreme sentiment from the list. For example, reviews containing extreme sentiment such as "This restaurant is amazing!!!" or "I'll never go there again."

[0500] The server then analyzes the reviewer's user profile. Using user profile analysis methods, the server analyzes the reviewer's posting frequency, geographic location, and the recency of the reviewer account. For example, it detects unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting in multiple locations. This is done using the database management systems MySQL and Apache Cassandra.

[0501] Finally, the server provides the filtered, trusted reviews to the user. When a user requests a review of a specific store or service from their device, the server sends only trusted reviews to the device. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0502] Examples:

[0503] A user uses a device to search for reviews of "restaurants in Tokyo."

[0504] The server collects reviews from an online review platform API.

[0505] The collected reviews are analyzed using natural language processing technology and an emotion engine to detect spam characteristics and extreme emotions.

[0506] The server filters out inappropriate reviews and analyzes user profiles to detect unusual activity.

[0507] Only trusted reviews resulting from the filtering are provided to the user.

[0508] Example prompt sentence:

[0509] "Write a program to identify users who submit a large number of reviews in a certain period of time and detect any unusual activity."

[0510] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0511] Step 1:

[0512] The server uses the API of the online review platform to collect review information. As input, it receives data such as the user ID, review content, rating, and posting date and time. Based on this, it stores the review information in a database. This is how the review information is collected.

[0513] Step 2:

[0514] The server analyzes the collected review information using a Python natural language processing library (e.g., NLTK or spaCy). The review content is given as input. First, the review text is preprocessed (tokenization, stop word removal), then spam filtering (Bayesian filtering, etc.) and grammatical consistency check (generating a parse tree). This allows the spam characteristics and grammatical consistency to be evaluated.

[0515] Step 3:

[0516] The server analyzes the sentiment of the reviews using a sentiment engine (e.g., a natural language sentiment API). The preprocessed review content is given as input. The sentiment engine analyzes the frequency and context of specific keywords and phrases, and outputs a sentiment score as positive, negative, or neutral. This classifies the sentiment of each review.

[0517] Step 4:

[0518] The server identifies and filters out inappropriate reviews. It receives inputs such as sentiment scores, spam characteristics, and grammatical consistency assessment results. The inappropriate review filtering method filters out reviews with spam characteristics or extreme sentiment, eliminating unreliable reviews.

[0519] Step 5:

[0520] The server analyzes the reviewer's user profile. As input, it takes information such as the reviewer's posting frequency, geographic location, and account recency. It uses a database management system (e.g., MySQL or Apache Cassandra) to detect unusual activity, such as users posting a large number of reviews in a short period of time or users posting reviews frequently in different locations. This allows it to assess the reviewer's trustworthiness.

[0521] Step 6:

[0522] A user requests reviews of a specific store or service using a device. A search keyword (e.g., "restaurants in Tokyo") is given as input. The server then provides filtered, trusted reviews, allowing users to make decisions based on accurate information.

[0523] (Application example 2)

[0524] 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."

[0525] Conventional review systems often contain spam reviews or reviews with extreme emotional expressions, making it difficult for users to obtain reliable reviews. Furthermore, the credibility of reviews based on reviewer profiles and activity histories is often insufficient, negatively impacting users' judgments. Furthermore, there has been no consistent method to effectively filter these reviews and provide only reliable reviews.

[0526] 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 a data collection means, a natural language processing means, an inappropriate review exclusion means, a user profile analysis means, a review providing means, a sentiment analysis means, and an application means to be installed on the terminal. This makes it possible to analyze review information collected from an online review platform, exclude spam and inappropriate reviews, and display only reliable reviews on the user's terminal.

[0527] "Data Collection Instruments" means instruments capable of collecting review information from online review platforms.

[0528] "Natural language processing means" refers to means for analyzing collected review information and performing spam characteristics, grammatical consistency, and sentiment analysis.

[0529] "Measures to filter out inappropriate reviews" are measures to detect and filter out reviews that have spam-like characteristics, reviews that contain extreme emotional expressions, and reviews with unnatural grammar.

[0530] "User profile analysis means" refers to means for analyzing profile information such as reviewer posting frequency, geographic location, and account recency.

[0531] The "review providing means" is a means for providing filtered and reliable review information to users.

[0532] The "sentiment analysis means" is a means for analyzing the content of a review and identifying positive, negative, or neutral sentiment.

[0533] The "application means installed on the terminal" refers to a means including an application that provides only reliable reviews when review information is displayed on the terminal.

[0534] This invention relates to a system that reliably filters review information collected from online review platforms and provides it to users. This system mainly consists of three elements: a server, a terminal, and a user.

[0535] The server collects review information from online review platforms. Specifically, the server uses a specific API to obtain review information from various review platforms. The collected data includes user IDs, review content, rating points, posting date and time, etc.

[0536] The server analyzes the collected review information using natural language processing. Utilizing natural language processing technology, the review content is subjected to spam filtering, grammar consistency checks, and sentiment analysis. This analysis determines whether each review is spam or contains unnatural content.

[0537] The server then uses sentiment analysis to analyze the review content and identify positive, negative, or neutral sentiment. Sentiment analysis is performed by analyzing the frequency of specific keywords and phrases, as well as contextual sentiment trends. During this process, if a review is extremely positive or negative, it is identified as spam or unnatural.

[0538] The server then detects and filters out inappropriate reviews, such as spam, reviews with extreme emotional expressions, and reviews with poor grammar, eliminating reviews with low credibility from the list.

[0539] The server also analyzes reviewer profiles using user profile analysis tools, which analyze reviewer posting frequency, geographic location, account recency, etc., to detect unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting from multiple different locations.

[0540] The terminal displays reviews of products and services searched for by the user through an application means installed on the terminal. The application has a function of displaying only reliable reviews sent from the server, allowing the user to make decisions based on reliable reviews.

[0541] For example, a user searches for a smartphone case on an online shopping site using a device. The server collects review information and analyzes the reviews using sentiment analysis and natural language processing. Inappropriate reviews are filtered out, and only reliable reviews are provided to the user's device after analyzing the user profile.

[0542] The program to realize this process is implemented using Python, Google Cloud Natural Language API, IBM Watson Tone Analyzer API, etc. This system enables users to make accurate decisions based on reliable reviews, and also enables companies and service providers to receive fair evaluations.

[0543] Example prompt: "Create a Python program that uses the data obtained from the review collection API call to analyze the review content using the Google Cloud Natural Language API and IBM Watson Tone Analyzer API, filter out spam and inappropriate reviews, and provide only trustworthy reviews to users."

[0544] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0545] Step 1:

[0546] Collecting review information

[0547] Subject: Server

[0548] Description: The server uses the API of an online review platform to collect review information for a specific product. For example, to obtain reviews for a specific product, the server calls the API and obtains data including the review content, user ID, rating, posting date and time, etc.

[0549] Input: Product ID given to the API

[0550] Output: A list of retrieved review information

[0551] Step 2:

[0552] Review analysis using natural language processing

[0553] Subject: Server

[0554] Description: The server analyzes the collected review information using a natural language processing engine (e.g., Google Cloud Natural Language API). Specifically, it evaluates the credibility of each review by filtering spam, checking grammar consistency, and analyzing sentiment.

[0555] Input: List of review information

[0556] Output: A list of parsed review information (with spam characteristics, grammar consistency, and sentiment scores)

[0557] Step 3:

[0558] Emotion analysis

[0559] Subject: Server

[0560] Description: The server uses a sentiment analysis engine to classify review content as positive, negative, or neutral. This process involves evaluating the frequency of specific keywords and phrases, the positive / negative tendencies of the context, and other factors to generate a sentiment score. Specifically, it uses the IBM Watson Tone Analyzer API.

[0561] Input: A list of parsed review information (with spam characteristics and grammar consistency)

[0562] Output: A list of reviews with sentiment scores

[0563] Step 4:

[0564] Filtering out inappropriate reviews

[0565] Subject: Server

[0566] Description: The server detects and filters out inappropriate reviews based on the sentiment score obtained by the sentiment analysis method. Reviews that are excessively positive or negative, spam-like, or have poor grammar are removed.

[0567] Input: A list of reviews with sentiment scores

[0568] Output: A list of filtered reviews

[0569] Step 5:

[0570] User profile analysis

[0571] Subject: Server

[0572] Description: The server analyzes the reviewer's user profile, which includes posting frequency, geographic location, account recency, etc. If a reviewer posts a large number of reviews in a short period of time or frequently from multiple different locations, this is detected as unusual activity.

[0573] Input: A filtered list of reviews

[0574] Output: A list of reliable reviews

[0575] Step 6:

[0576] Providing a review

[0577] Subject: Terminal

[0578] Description: Provide users with reliable reviews through an application installed on a terminal, which displays reliable review information received from a server to the user.

[0579] Input: A list of reliable reviews

[0580] Output: Reliable review information provided to users

[0581] In this way, the server and the terminal work together to realize a system that provides highly reliable review information to users.

[0582] 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.

[0583] 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.

[0584] 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.

[0585] [Third embodiment]

[0586] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0587] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0588] 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).

[0589] 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.

[0590] 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.

[0591] 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).

[0592] 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.

[0593] 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.

[0594] 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.

[0595] 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.

[0596] 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.

[0597] 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."

[0598] This invention is a system for reliably filtering review information collected from an online review platform. The system operates as follows between a server, a terminal, and a user.

[0599] First, the server collects review information from online review platforms. For example, it uses APIs from Google Maps or other review platforms to obtain reviews about specified stores or services. The collected review information is then stored as a dataset.

[0600] The server then analyzes the collected reviews using natural language processing technology. Specifically, it uses a natural language processing library to analyze the text of each review and perform spam characteristics, grammatical consistency, sentiment analysis, etc. This analysis provides information such as whether each review is spam or contains unnatural content.

[0601] Next, the server filters out inappropriate reviews. This filter removes spam-like reviews, unnatural grammar, or extreme emotional expressions from the list. For example, it filters out reviews that contain the same text repeatedly or that are posted in large numbers in a short period of time.

[0602] The server then analyzes the reviewer's user profile to detect whether the reviewer is engaging in unusual activity, such as multiple reviews submitted by the same user in a short period of time, or frequent reviews submitted in a particular location. Reviews submitted by reviewers with such unusual activity are also filtered out.

[0603] Finally, the server provides the filtered, reliable reviews to the user. When a user requests a review of a specific store or service from their device, the server returns the filtered reviews. This allows users to make decisions based on reliable reviews, and companies and stores can also receive fair evaluations.

[0604] Examples:

[0605] A user uses a device to search for reviews of "restaurants in Tokyo."

[0606] The server collects reviews from Google Maps and other review platforms.

[0607] The collected reviews are analyzed using natural language processing technology to detect spam characteristics and unnatural content.

[0608] The server filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual activity.

[0609] Only trusted reviews resulting from the filtering are provided to the user.

[0610] This is how the system can be implemented. This process allows consumers to make informed decisions and businesses to receive fair reviews.

[0611] The processing flow will be explained below.

[0612] Step 1:

[0613] The server collects review information from online review platforms. Specifically, the server sends a request to the API of Google Maps or other review platforms to retrieve reviews of the specified store or service. The review information returned from the API includes details such as the user ID, review content, rating, and posting date and time.

[0614] Step 2:

[0615] The reviews collected by the server are analyzed using natural language processing. Specifically, a natural language processing library (e.g., spaCy or NLTK) is used to analyze the review text and perform spam characteristics, grammatical consistency, and sentiment analysis. Spam characteristics include analyzing repeated phrases, excessive use of links, and non-natural language. Grammar consistency detects grammatical errors and inappropriate phrases, and sentiment analysis evaluates the review's positive, negative, or neutral sentiment.

[0616] Step 3:

[0617] The server identifies and filters out inappropriate reviews. Using inappropriate review filtering methods, it removes reviews with spam characteristics, reviews with extremely poor grammar, and reviews containing extreme emotional expressions from the list. For example, reviews with excessive emotional expressions like "This store is great!!!" and extremely negative reviews like "It was a terrible experience" are filtered out.

[0618] Step 4:

[0619] The server analyzes the reviewer's user profile. It uses user profile analysis tools to analyze the reviewer's profile. Specifically, it considers the reviewer's posting frequency, the geographic location of the posts, and whether the reviewer's account is newly created. This analysis identifies unusual activity when the same user posts a large number of reviews in a short period of time or frequently posts from multiple different locations.

[0620] Step 5:

[0621] The server provides users with filtered, reliable reviews. When a user requests reviews for a specific store or service from their device, the server returns only reliable reviews that have gone through the filtering process described above. For example, when a user searches for reviews for "restaurants in Tokyo," the server collects, analyzes, and filters reviews from Google Maps and other review platforms and displays them to the user.

[0622] These are the processing steps of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0623] Example 1

[0624] 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."

[0625] Conventional online review platforms contain a large number of spam and inappropriate reviews, making it difficult for users to make decisions based on accurate and reliable information. Furthermore, the lack of mechanisms to detect abnormal reviewer activity has led to issues with impartial evaluations.

[0626] 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.

[0627] In this invention, the server includes a data collection unit, a natural language processing unit, an inappropriate review exclusion unit, a user profile analysis unit, and a review providing unit. This allows for a consistent process of review information collection, analysis, filtering, and provision. Specifically, review information is collected from an online review platform and analyzed using natural language processing technology. Furthermore, inappropriate reviews are identified and excluded using a machine learning algorithm, and abnormal reviewer activity is detected using a clustering algorithm, thereby providing users with reliable review information.

[0628] "Data collection means" refers to a function that automatically obtains review information from online review platforms.

[0629] "Natural language processing means" is a technology that analyzes collected review information and performs tasks such as spam characteristics, grammatical consistency, and sentiment analysis.

[0630] "Means for filtering out inappropriate reviews" is a function that identifies and filters out reviews that contain spam or unnatural language from the collected review information.

[0631] "User profile analysis means" is a function for analyzing reviewer activity patterns and detecting unnatural behavior.

[0632] The "review providing means" is a function that provides filtered, reliable review information in a format that is accessible to users.

[0633] An "online review platform" is a system that allows users to post and view reviews of products and services on the Internet.

[0634] "Machine learning algorithm" is a general term for programs and techniques that automatically learn patterns from data and apply them to new data.

[0635] A "clustering algorithm" is a technique for dividing data into groups, and is a method used in particular to detect anomalous data points.

[0636] "Analysis" is the process of processing collected information for a specific purpose to reveal its characteristics and trends.

[0637] "Filtering" is the act of selecting data based on specific conditions and eliminating unnecessary information.

[0638] The present invention provides a system for reliably filtering review information collected from online review platforms, including a data collection unit, a natural language processing unit, a unit for filtering inappropriate reviews, a unit for analyzing user profiles, and a unit for providing reviews.

[0639] First, the server uses data collection means to collect review information from online review platforms. Specifically, it uses APIs to obtain reviews of designated stores and services from Google Maps and other review platforms. It then secures appropriate access rights using API keys and OAuth authentication, and stores the collected review information in a database.

[0640] The server then analyzes the collected review text using natural language processing (NLP) libraries such as Python's NLTK and spaCy, which determine each review's spam characteristics, grammatical consistency, and sentiment analysis to assess its trustworthiness.

[0641] The server then runs an inappropriate review filtering process, which uses machine learning algorithms (e.g., XGBoost) to automatically identify and filter out spam and unnatural language from the analyzed reviews, ensuring that users only see trustworthy reviews.

[0642] The server then uses user profile analysis tools to analyze the reviewer's activity. Clustering algorithms (e.g., DBSCAN) are used to detect anomalous behavior, such as reviewers posting a large number of reviews in a short period of time or frequently posting reviews in a particular location. Reviews from reviewers with abnormal activity are also rejected.

[0643] Finally, the server provides filtered and reliable reviews to users through a review providing means. When a user requests a review of a specific store or service from their device, the server can return only reliable reviews. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0644] As a concrete example, consider a case where a user searches for reviews of "restaurants in Tokyo." The server collects reviews from Google Maps and other review platforms and analyzes them using natural language processing technology. It then filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual behavior. Finally, the filtered, reliable reviews are provided to the user.

[0645] Example prompt sentence:

[0646] "Show me restaurant reviews in Tokyo."

[0647] This process allows users to make decisions based on accurate and reliable information.

[0648] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0649] Step 1:

[0650] The server starts collecting data. It uses the API of the online review platform to obtain review information about the specified store or service. It receives the store or service ID, API key, and authentication information as input, and generates JSON data of the obtained review information as output. This data is then stored in a database.

[0651] Step 2:

[0652] The server applies natural language processing technology. It uses natural language processing libraries (such as NLTK and spaCy) to analyze the collected review information (text data). It takes the review text data as input and outputs the results of spam characteristics, grammatical consistency, and sentiment analysis. Specifically, it performs sentence tokenization, morphological analysis, and sentiment score calculation.

[0653] Step 3:

[0654] The server filters out inappropriate reviews. It uses machine learning algorithms (such as XGBoost) to classify reviews with spam characteristics or unnatural grammar as inappropriate reviews. It takes the results of natural language processing as input and identifies and filters out reviews with scores that are deemed inappropriate. The output is a filtered list of appropriate reviews.

[0655] Step 4:

[0656] The server analyzes the reviewer's user profile, examines the reviewer's posting frequency and activity patterns, and uses a clustering algorithm (such as DBSCAN) to detect fraudulent activity. It takes the reviewer's posting data (posting date and time, location, etc.) as input and filters out reviews from reviewers who are determined to be anomalous. The output is a list of highly reliable reviews that have been further filtered.

[0657] Step 5:

[0658] The server provides the filtered reviews to the user. When a user requests reviews of a specific store or service from their device, the server returns highly reliable reviews. The user inputs the ID of the store or service in question, and the output is highly reliable review information. This allows users to make decisions based on accurate and reliable information.

[0659] Through the above steps, the system can reliably filter review information collected from online review platforms and provide accurate information to users.

[0660] (Application example 1)

[0661] 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."

[0662] Some reviews provided on online review platforms are inappropriate or unreliable, making it difficult for users to make decisions based on accurate information. This problem is particularly pronounced when purchasing products on e-commerce platforms, where users need to make purchasing decisions based on unbiased and reliable reviews.

[0663] 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.

[0664] In this invention, the server includes a data collection means, a natural language processing means, an inappropriate review exclusion means, a user profile analysis means, a reliable review providing means, a product identification means, and a review data display means, thereby providing reliable reviews to users and enabling users to make decisions based on more accurate information.

[0665] "Data Collection Means" means means for collecting review information from online review platforms or e-commerce platforms.

[0666] The "natural language processing means" is a means for analyzing collected reviews and performing spam characteristics, grammatical consistency, sentiment analysis, and context analysis.

[0667] "Inappropriate review filtering" refers to filtering out reviews that have spam characteristics or contain unnatural grammar or extreme emotional expressions.

[0668] "User profile analysis means" is a means of analyzing whether reviewers are engaging in unnatural activities.

[0669] A "means for providing reliable reviews" is a means for providing filtered, reliable reviews to users.

[0670] A "product identification means" is a means that allows a user to search for reviews about a particular product.

[0671] The "review data display means" is a means for visually displaying reliable reviews to the user.

[0672] This invention is a system that reliably filters review information collected from online review platforms or e-commerce platforms and provides users with the most reliable reviews. Implementing this system requires a server, a user terminal, and a program for linking them.

[0673] 1. System Program Overview

[0674] The server executes a program including a data collection means, a natural language processing means, a means for excluding inappropriate reviews, a means for analyzing user profiles, a means for providing reliable reviews, a means for identifying products, and a means for displaying review data. The program performs the following processes:

[0675] 2. Hardware and Software Used

[0676] Hardware:

[0677] Cloud server: The central hardware that collects and analyzes data.

[0678] Smartphone: The device where users search for reviews and view filtered reviews.

[0679] software:

[0680] Python: A programming language for natural language processing and data analysis.

[0681] TextBlob: A natural language processing library for performing sentiment analysis on reviews.

[0682] NLTK: A natural language processing library used to analyze text data.

[0683] Requests: An HTTP library for collecting review information through an API.

[0684] 3. Program processing explanation

[0685] The server uses the data collection means to collect review information from the online review platform or the e-commerce platform via an API, and the collected review information is stored on the server as a data set.

[0686] Next, we analyze the collected reviews using natural language processing tools. Specifically, we use TextBlob to analyze the sentiment and grammatical consistency of each review, and to detect spam characteristics. We also use NLTK for contextual analysis.

[0687] As a means of excluding inappropriate reviews, we remove reviews that have spam characteristics, reviews that lack grammar consistency, and reviews that contain extreme emotional expressions from the list. We also use user profile analysis to detect unusual reviewer activity and exclude such reviews.

[0688] Finally, the reliable review providing means provides filtered reliable reviews to the user's smartphone. When the user searches for a specific product, the product is identified by the product identification means and visually displayed by the review data display means.

[0689] 4. Examples of concrete examples and prompts

[0690] Example of a user purchasing an iPhone 12:

[0691] A user searches for "iPhone 12" reviews on their smartphone.

[0692] A server collects reviews from an online review platform.

[0693] The collected reviews are analyzed using natural language processing technology to determine spam characteristics, grammatical consistency, sentiment analysis, and context analysis.

[0694] Inappropriate or unreliable reviews will be filtered out.

[0695] Appropriate reviews are filtered and presented to users as trusted reviews.

[0696] Example prompt for a generative AI model:

[0697] Product being reviewed: iPhone 12

[0698] Review List:

[0699] 1. Review ID: 001

[0700] It says: "This product is great. I'm happy with the price and quality."

[0701] Sentiment analysis result: Positive polarity (0.8)

[0702] 2. Review ID: 002

[0703] Content: "spam, fake review"

[0704] Sentiment analysis results: Neutral (0)

[0705] 3. Review ID: 003

[0706] What it said: "Very good product, but a little pricey. Service was good."

[0707] Sentiment analysis result: Positive polarity (0.5)

[0708] This way, users can make decisions based on reliable reviews.

[0709] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0710] Step 1:

[0711] Data collection

[0712] The server uses the data collection means to obtain review information from online review platforms or e-commerce platforms via API. The input is the API endpoint of each platform, and the output is a dataset of collected reviews. Specifically, it sends an HTTP request and stores the obtained data in JSON format on the server.

[0713] Step 2:

[0714] Natural Language Processing

[0715] The server uses natural language processing tools to analyze the collected reviews. The input is a dataset of reviews obtained by the data collection tools, and the output is a dataset of analyzed reviews. Specifically, it uses the TextBlob library to analyze the text of each review and perform sentiment analysis (positive, negative, neutral), grammatical consistency checks, and spam detection.

[0716] Step 3:

[0717] Inappropriate review exclusion

[0718] The server uses inappropriate review filtering methods to filter out reviews with spam characteristics, unnatural grammar, or extreme emotional expressions. The input is a dataset of reviews analyzed using natural language processing methods, and the output is a dataset of reliable reviews. Specifically, it filters out reviews that have been detected as spam characteristics, grammatically incorrect reviews, and reviews with extremely biased emotions.

[0719] Step 4:

[0720] User profile analysis

[0721] The server uses user profile analysis to analyze reviewer activity and detect unusual activity. The input is a filtered review dataset, and the output is a more reliable review dataset. Specifically, it detects cases where a large number of reviews are posted by the same user in a short period of time, or where reviews are frequently posted in a specific location, and excludes those reviews.

[0722] Step 5:

[0723] Providing reliable reviews

[0724] The server uses the reliable review providing means to provide filtered reliable reviews to the user. The input is the final filtered review dataset, and the output is the reliable reviews displayed on the user's device. Specifically, based on the user's request, the server retrieves reliable reviews related to a specific product from the database and sends them in JSON format to the user's smartphone.

[0725] Step 6:

[0726] Product Identification

[0727] The user's device uses a product identification means to search for a specific product. The input is the product name and product ID searched by the user, and the output is a list of reviews related to that product. Specifically, when the user enters the product name into the smartphone app, the ID corresponding to that product is sent to the server.

[0728] Step 7:

[0729] Review data display

[0730] The user's device visually displays the reliable reviews using the review data display means. The input is a dataset of reliable reviews obtained from the server, and the output is review information displayed on the user's smartphone screen. Specifically, the application formats the obtained reviews and displays them in a format that is easy for the user to view.

[0731] The above processing steps allow users to select products with confidence based on highly reliable reviews.

[0732] 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.

[0733] This invention combines a system that reliably filters review information collected from online review platforms with an emotion engine that recognizes user emotions. The system operates as follows between a server, a terminal, and a user.

[0734] First, the server collects review information from online review platforms. Specifically, the server uses a specific API to obtain review data from Google Maps and other review platforms. The collected data includes user IDs, review content, rating points, posting date and time, etc.

[0735] The server then analyzes the collected review information using natural language processing technology. Using natural language processing techniques, the review text undergoes spam filtering, grammar consistency checks, and sentiment analysis. This analysis identifies whether each review is spam or contains unnatural content.

[0736] The server then uses a sentiment engine to analyze the text data in the review and identify positive, negative, or neutral sentiment. The sentiment engine evaluates the review's sentiment by analyzing the frequency of specific keywords and phrases, the positive / negative tendencies of the context, and other factors. During this process, if a review is very positive or very negative, it is identified as spam or unnatural.

[0737] Next, the server identifies and filters out inappropriate reviews. Inappropriate review filtering measures remove reviews with spam characteristics, poor grammar, or extreme sentiment from the list. For example, reviews containing extreme sentiment such as "This restaurant is amazing!!!" or "I'll never go there again."

[0738] The server then analyzes the reviewer's user profile, using user profile analysis tools to analyze the reviewer's posting frequency, geographic location, and the recency of the reviewer account, among other things. This analysis detects unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting in multiple locations.

[0739] Finally, the server provides users with filtered, reliable reviews. When a user requests a review of a specific store or service from their device, the server provides only reliable reviews. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0740] Examples:

[0741] A user uses a device to search for reviews of "restaurants in Tokyo."

[0742] The server collects reviews from Google Maps and other review platforms.

[0743] The collected reviews are analyzed using natural language processing technology and an emotion engine to detect spam characteristics and extreme emotions.

[0744] The server filters out inappropriate reviews and analyzes user profiles to detect unusual activity.

[0745] Only trusted reviews resulting from the filtering are provided to the user.

[0746] This is the implementation form of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0747] The processing flow will be explained below.

[0748] Step 1:

[0749] The server collects review information from online review platforms. Specifically, it uses the APIs of Google Maps and other review platforms to obtain reviews of specified stores and services. This review information includes user IDs, review content, ratings, posting dates, etc.

[0750] Step 2:

[0751] The server analyzes the collected reviews using natural language processing. Specifically, it uses a natural language processing library to analyze the text data and perform spam characteristics, grammatical consistency, and sentiment analysis. For example, it detects the presence of specific spam keywords and grammatical errors, and determines whether the sentiment of the comment is positive, negative, or neutral.

[0752] Step 3:

[0753] The server uses a sentiment engine to evaluate the detailed sentiment of the analyzed reviews. The sentiment engine analyzes specific keywords and phrases within the review to generate an overall sentiment score. Based on this score, the server identifies whether the review is excessively positive or negative. Based on this result, the server can determine whether an overly emotional review is inappropriate.

[0754] Step 4:

[0755] The server identifies and filters out inappropriate reviews. It uses inappropriate review filtering to remove reviews that have spammy characteristics, poor grammar, or extreme sentiments that are identified by the sentiment engine. For example, reviews like "This store is great!!!" and "I'll never go there again" are filtered out.

[0756] Step 5:

[0757] The server analyzes the reviewer's user profile. Using user profile analysis methods, it checks the reviewer's posting frequency, geographic location, account recency, etc. This profile analysis identifies unusual activity when the same user posts a large number of reviews in a short period of time or frequently posts reviews in different locations.

[0758] Step 6:

[0759] The server provides users with filtered, reliable reviews. When a user requests reviews for a specific store or service using their device, the server returns only the filtered, reliable reviews, allowing users to make decisions based on accurate information.

[0760] Examples:

[0761] A user uses a device to search for reviews of "restaurants in Tokyo."

[0762] The server collects review information from Google Maps and other review platforms.

[0763] The collected reviews are analyzed using natural language processing tools and an emotion engine to detect spam characteristics and emotional extremes.

[0764] The server filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual behavior.

[0765] The server provides filtered and trusted reviews to the user.

[0766] These are the specific processing steps of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0767] Example 2

[0768] 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."

[0769] Online review platforms are often filled with reviews containing unreliable information, spam reviews, and unnatural emotional expressions. The presence of such inappropriate reviews makes it difficult for users to make accurate decisions. Furthermore, reviewer credibility is not evaluated, and inappropriate activity may be overlooked. Therefore, a system is needed that provides reliable reviews and enables users to make accurate decisions.

[0770] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0771] In this invention, the server includes a data collection means, a natural language processing means, a sentiment analysis means, an inappropriate review excluding means, a user profile analysis means, and a review providing means, thereby enabling the server to reliably filter review information collected from online review platforms and provide accurate and fair reviews to users.

[0772] "Data collection means" refers to a device that has the function of collecting review information from an online review platform.

[0773] The "natural language processing means" is a device that has the function of analyzing collected reviews and performing spam characteristics, grammatical consistency, and sentiment analysis.

[0774] The "sentiment analysis means" is a device that has the function of analyzing text data in reviews and identifying positive, negative, and neutral emotions.

[0775] An "inappropriate review filtering method" is a device that has the function of excluding reviews that have spam characteristics, poor grammar, or extreme emotions from the list.

[0776] A "user profile analysis means" is a device that has the function of analyzing a reviewer's posting frequency, geographic location, the recency of the reviewer account, etc., and detecting any unusual activity.

[0777] The "review providing means" is a device that has the function of providing filtered and reliable reviews to users.

[0778] This invention combines a system for reliably filtering review information collected from online review platforms with an emotion engine that recognizes user emotions. The system operates among a server, a terminal, and a user as follows: The server processes and calculates data using specific hardware and software.

[0779] First, the server collects review information from an online review platform. Specifically, the server obtains review data using a specific API (e.g., online review platform API). The collected data includes user IDs, review content, rating points, posting date and time, etc.

[0780] The server then analyzes the collected review information using natural language processing technology. The natural language processing uses Python natural language processing libraries such as NLTK and spaCy. The server then performs spam filtering, grammar consistency checks, and sentiment analysis. This allows it to determine whether each review is spam or contains unnatural content.

[0781] Additionally, the server uses a sentiment engine (e.g., natural language sentiment API) to analyze the text data in the review and identify positive, negative, or neutral sentiment. The sentiment engine evaluates the sentiment of the review by analyzing the frequency of specific keywords and phrases, the positive / negative tendencies of the context, etc. During this process, if a review is very positive or very negative, it is identified as spam or unnatural.

[0782] Next, the server identifies and filters out inappropriate reviews. Inappropriate review filtering measures remove reviews with spam characteristics, poor grammar, or extreme sentiment from the list. For example, reviews containing extreme sentiment such as "This restaurant is amazing!!!" or "I'll never go there again."

[0783] The server then analyzes the reviewer's user profile. Using user profile analysis methods, the server analyzes the reviewer's posting frequency, geographic location, and the recency of the reviewer account. For example, it detects unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting in multiple locations. This is done using the database management systems MySQL and Apache Cassandra.

[0784] Finally, the server provides the filtered, trusted reviews to the user. When a user requests a review of a specific store or service from their device, the server sends only trusted reviews to the device. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0785] Examples:

[0786] A user uses a device to search for reviews of "restaurants in Tokyo."

[0787] The server collects reviews from an online review platform API.

[0788] The collected reviews are analyzed using natural language processing technology and an emotion engine to detect spam characteristics and extreme emotions.

[0789] The server filters out inappropriate reviews and analyzes user profiles to detect unusual activity.

[0790] Only trusted reviews resulting from the filtering are provided to the user.

[0791] Example prompt sentence:

[0792] "Write a program to identify users who submit a large number of reviews in a certain period of time and detect any unusual activity."

[0793] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0794] Step 1:

[0795] The server uses the API of the online review platform to collect review information. As input, it receives data such as the user ID, review content, rating, and posting date and time. Based on this, it stores the review information in a database. This is how the review information is collected.

[0796] Step 2:

[0797] The server analyzes the collected review information using a Python natural language processing library (e.g., NLTK or spaCy). The review content is given as input. First, the review text is preprocessed (tokenization, stop word removal), then spam filtering (Bayesian filtering, etc.) and grammatical consistency check (generating a parse tree). This allows the spam characteristics and grammatical consistency to be evaluated.

[0798] Step 3:

[0799] The server analyzes the sentiment of the reviews using a sentiment engine (e.g., a natural language sentiment API). The preprocessed review content is given as input. The sentiment engine analyzes the frequency and context of specific keywords and phrases, and outputs a sentiment score as positive, negative, or neutral. This classifies the sentiment of each review.

[0800] Step 4:

[0801] The server identifies and filters out inappropriate reviews. It receives inputs such as sentiment scores, spam characteristics, and grammatical consistency assessment results. The inappropriate review filtering method filters out reviews with spam characteristics or extreme sentiment, eliminating unreliable reviews.

[0802] Step 5:

[0803] The server analyzes the reviewer's user profile. As input, it takes information such as the reviewer's posting frequency, geographic location, and account recency. It uses a database management system (e.g., MySQL or Apache Cassandra) to detect unusual activity, such as users posting a large number of reviews in a short period of time or users posting reviews frequently in different locations. This allows it to assess the reviewer's trustworthiness.

[0804] Step 6:

[0805] A user requests reviews of a specific store or service using a device. A search keyword (e.g., "restaurants in Tokyo") is given as input. The server then provides filtered, trusted reviews, allowing users to make decisions based on accurate information.

[0806] (Application example 2)

[0807] 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."

[0808] Conventional review systems often contain spam reviews or reviews with extreme emotional expressions, making it difficult for users to obtain reliable reviews. Furthermore, the credibility of reviews based on reviewer profiles and activity histories is often insufficient, negatively impacting users' judgments. Furthermore, there has been no consistent method to effectively filter these reviews and provide only reliable reviews.

[0809] 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 a data collection means, a natural language processing means, an inappropriate review exclusion means, a user profile analysis means, a review providing means, a sentiment analysis means, and an application means to be installed on the terminal. This makes it possible to analyze review information collected from an online review platform, exclude spam and inappropriate reviews, and display only reliable reviews on the user's terminal.

[0810] "Data Collection Instruments" means instruments capable of collecting review information from online review platforms.

[0811] "Natural language processing means" refers to means for analyzing collected review information and performing spam characteristics, grammatical consistency, and sentiment analysis.

[0812] "Measures to filter out inappropriate reviews" are measures to detect and filter out reviews that have spam-like characteristics, reviews that contain extreme emotional expressions, and reviews with unnatural grammar.

[0813] "User profile analysis means" refers to means for analyzing profile information such as reviewer posting frequency, geographic location, and account recency.

[0814] The "review providing means" is a means for providing filtered and reliable review information to users.

[0815] The "sentiment analysis means" is a means for analyzing the content of a review and identifying positive, negative, or neutral sentiment.

[0816] The "application means installed on the terminal" refers to a means including an application that provides only reliable reviews when review information is displayed on the terminal.

[0817] This invention relates to a system that reliably filters review information collected from online review platforms and provides it to users. This system mainly consists of three elements: a server, a terminal, and a user.

[0818] The server collects review information from online review platforms. Specifically, the server uses a specific API to obtain review information from various review platforms. The collected data includes user IDs, review content, rating points, posting date and time, etc.

[0819] The server analyzes the collected review information using natural language processing. Utilizing natural language processing technology, the review content is subjected to spam filtering, grammar consistency checks, and sentiment analysis. This analysis determines whether each review is spam or contains unnatural content.

[0820] The server then uses sentiment analysis to analyze the review content and identify positive, negative, or neutral sentiment. Sentiment analysis is performed by analyzing the frequency of specific keywords and phrases, as well as contextual sentiment trends. During this process, if a review is extremely positive or negative, it is identified as spam or unnatural.

[0821] The server then detects and filters out inappropriate reviews, such as spam, reviews with extreme emotional expressions, and reviews with poor grammar, eliminating reviews with low credibility from the list.

[0822] The server also analyzes reviewer profiles using user profile analysis tools, which analyze reviewer posting frequency, geographic location, account recency, etc., to detect unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting from multiple different locations.

[0823] The terminal displays reviews of products and services searched for by the user through an application means installed on the terminal. The application has a function of displaying only reliable reviews sent from the server, allowing the user to make decisions based on reliable reviews.

[0824] For example, a user searches for a smartphone case on an online shopping site using a device. The server collects review information and analyzes the reviews using sentiment analysis and natural language processing. Inappropriate reviews are filtered out, and only reliable reviews are provided to the user's device after analyzing the user profile.

[0825] The program to realize this process is implemented using Python, Google Cloud Natural Language API, IBM Watson Tone Analyzer API, etc. This system enables users to make accurate decisions based on reliable reviews, and also enables companies and service providers to receive fair evaluations.

[0826] Example prompt: "Create a Python program that uses the data obtained from the review collection API call to analyze the review content using the Google Cloud Natural Language API and IBM Watson Tone Analyzer API, filter out spam and inappropriate reviews, and provide only trustworthy reviews to users."

[0827] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0828] Step 1:

[0829] Collecting review information

[0830] Subject: Server

[0831] Description: The server uses the API of an online review platform to collect review information for a specific product. For example, to obtain reviews for a specific product, the server calls the API and obtains data including the review content, user ID, rating, posting date and time, etc.

[0832] Input: Product ID given to the API

[0833] Output: A list of retrieved review information

[0834] Step 2:

[0835] Review analysis using natural language processing

[0836] Subject: Server

[0837] Description: The server analyzes the collected review information using a natural language processing engine (e.g., Google Cloud Natural Language API). Specifically, it evaluates the credibility of each review by filtering spam, checking grammar consistency, and analyzing sentiment.

[0838] Input: List of review information

[0839] Output: A list of parsed review information (with spam characteristics, grammar consistency, and sentiment scores)

[0840] Step 3:

[0841] Emotion analysis

[0842] Subject: Server

[0843] Description: The server uses a sentiment analysis engine to distinguish between positive, negative, and neutral review content. This process involves evaluating the frequency of specific keywords and phrases, the positive / negative tendencies of the context, and other factors to generate a sentiment score. Specifically, it uses the IBM Watson Tone Analyzer API.

[0844] Input: A list of parsed review information (with spam characteristics and grammar consistency)

[0845] Output: A list of reviews with sentiment scores

[0846] Step 4:

[0847] Filtering out inappropriate reviews

[0848] Subject: Server

[0849] Description: The server detects and filters out inappropriate reviews based on the sentiment score obtained by the sentiment analysis method. Reviews that are excessively positive or negative, spam-like, or have poor grammar are removed.

[0850] Input: A list of reviews with sentiment scores

[0851] Output: A list of filtered reviews

[0852] Step 5:

[0853] User profile analysis

[0854] Subject: Server

[0855] Description: The server analyzes the reviewer's user profile, which includes posting frequency, geographic location, account recency, etc. If a reviewer posts a large number of reviews in a short period of time or frequently from multiple different locations, this is detected as unusual activity.

[0856] Input: A filtered list of reviews

[0857] Output: A list of reliable reviews

[0858] Step 6:

[0859] Providing a review

[0860] Subject: Terminal

[0861] Description: Provide users with reliable reviews through an application installed on a terminal, which displays reliable review information received from a server to the user.

[0862] Input: A list of reliable reviews

[0863] Output: Reliable review information provided to users

[0864] In this way, the server and the terminal work together to realize a system that provides highly reliable review information to users.

[0865] 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.

[0866] 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.

[0867] 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.

[0868] [Fourth embodiment]

[0869] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[0870] 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.

[0871] 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).

[0872] 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.

[0873] 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.

[0874] 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).

[0875] 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.

[0876] 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.

[0877] 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.

[0878] 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.

[0879] 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.

[0880] 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.

[0881] 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."

[0882] This invention is a system for reliably filtering review information collected from an online review platform. The system operates as follows between a server, a terminal, and a user.

[0883] First, the server collects review information from online review platforms. For example, it uses APIs from Google Maps or other review platforms to obtain reviews about specified stores or services. The collected review information is then stored as a dataset.

[0884] The server then analyzes the collected reviews using natural language processing technology. Specifically, it uses a natural language processing library to analyze the text of each review and perform spam characteristics, grammatical consistency, sentiment analysis, etc. This analysis provides information such as whether each review is spam or contains unnatural content.

[0885] Next, the server filters out inappropriate reviews. This filter removes spam-like reviews, unnatural grammar, or extreme emotional expressions from the list. For example, it filters out reviews that contain the same text repeatedly or that are posted in large numbers in a short period of time.

[0886] The server then analyzes the reviewer's user profile to detect whether the reviewer is engaging in unusual activity, such as multiple reviews submitted by the same user in a short period of time, or frequent reviews submitted in a particular location. Reviews submitted by reviewers with such unusual activity are also filtered out.

[0887] Finally, the server provides the filtered, reliable reviews to the user. When a user requests a review of a specific store or service from their device, the server returns the filtered reviews. This allows users to make decisions based on reliable reviews, and companies and stores can also receive fair evaluations.

[0888] Examples:

[0889] A user uses a device to search for reviews of "restaurants in Tokyo."

[0890] The server collects reviews from Google Maps and other review platforms.

[0891] The collected reviews are analyzed using natural language processing technology to detect spam characteristics and unnatural content.

[0892] The server filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual activity.

[0893] Only trusted reviews resulting from the filtering are provided to the user.

[0894] This is how the system can be implemented. This process allows consumers to make informed decisions and businesses to receive fair reviews.

[0895] The processing flow will be explained below.

[0896] Step 1:

[0897] The server collects review information from online review platforms. Specifically, the server sends a request to the API of Google Maps or other review platforms to retrieve reviews of the specified store or service. The review information returned from the API includes details such as the user ID, review content, rating, and posting date and time.

[0898] Step 2:

[0899] The reviews collected by the server are analyzed using natural language processing. Specifically, a natural language processing library (e.g., spaCy or NLTK) is used to analyze the review text and perform spam characteristics, grammatical consistency, and sentiment analysis. Spam characteristics include analyzing repeated phrases, excessive use of links, and non-natural language. Grammar consistency detects grammatical errors and inappropriate phrases, and sentiment analysis evaluates the review's positive, negative, or neutral sentiment.

[0900] Step 3:

[0901] The server identifies and filters out inappropriate reviews. Using inappropriate review filtering methods, it removes reviews with spam characteristics, reviews with extremely poor grammar, and reviews containing extreme emotional expressions from the list. For example, reviews with excessive emotional expressions like "This store is great!!!" and extremely negative reviews like "It was a terrible experience" are filtered out.

[0902] Step 4:

[0903] The server analyzes the reviewer's user profile. It uses user profile analysis tools to analyze the reviewer's profile. Specifically, it considers the reviewer's posting frequency, the geographic location of the posts, and whether the reviewer's account is newly created. This analysis identifies unusual activity when the same user posts a large number of reviews in a short period of time or frequently posts from multiple different locations.

[0904] Step 5:

[0905] The server provides users with filtered, reliable reviews. When a user requests reviews for a specific store or service from their device, the server returns only reliable reviews that have gone through the filtering process described above. For example, when a user searches for reviews for "restaurants in Tokyo," the server collects, analyzes, and filters reviews from Google Maps and other review platforms and displays them to the user.

[0906] These are the processing steps of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[0907] Example 1

[0908] 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."

[0909] Conventional online review platforms contain a large number of spam and inappropriate reviews, making it difficult for users to make decisions based on accurate and reliable information. Furthermore, the lack of mechanisms to detect abnormal reviewer activity has led to issues with impartial evaluations.

[0910] 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.

[0911] In this invention, the server includes a data collection unit, a natural language processing unit, an inappropriate review exclusion unit, a user profile analysis unit, and a review providing unit. This allows for a consistent process of review information collection, analysis, filtering, and provision. Specifically, review information is collected from an online review platform and analyzed using natural language processing technology. Furthermore, inappropriate reviews are identified and excluded using a machine learning algorithm, and abnormal reviewer activity is detected using a clustering algorithm, thereby providing users with reliable review information.

[0912] "Data collection means" refers to a function that automatically obtains review information from online review platforms.

[0913] "Natural language processing means" is a technology that analyzes collected review information and performs tasks such as spam characteristics, grammatical consistency, and sentiment analysis.

[0914] "Means for filtering out inappropriate reviews" is a function that identifies and filters out reviews that contain spam or unnatural language from the collected review information.

[0915] "User profile analysis means" is a function for analyzing reviewer activity patterns and detecting unnatural behavior.

[0916] The "review providing means" is a function that provides filtered, reliable review information in a format that is accessible to users.

[0917] An "online review platform" is a system that allows users to post and view reviews of products and services on the Internet.

[0918] "Machine learning algorithm" is a general term for programs and techniques that automatically learn patterns from data and apply them to new data.

[0919] A "clustering algorithm" is a technique for dividing data into groups, and is a method used in particular to detect anomalous data points.

[0920] "Analysis" is the process of processing collected information for a specific purpose to reveal its characteristics and trends.

[0921] "Filtering" is the act of selecting data based on specific conditions and eliminating unnecessary information.

[0922] The present invention provides a system for reliably filtering review information collected from online review platforms, including a data collection unit, a natural language processing unit, a unit for filtering inappropriate reviews, a unit for analyzing user profiles, and a unit for providing reviews.

[0923] First, the server uses data collection means to collect review information from online review platforms. Specifically, it uses APIs to obtain reviews of designated stores and services from Google Maps and other review platforms. It then secures appropriate access rights using API keys and OAuth authentication, and stores the collected review information in a database.

[0924] The server then analyzes the collected review text using natural language processing (NLP) libraries such as Python's NLTK and spaCy, which determine each review's spam characteristics, grammatical consistency, and sentiment analysis to assess its trustworthiness.

[0925] The server then runs an inappropriate review filtering process, which uses machine learning algorithms (e.g., XGBoost) to automatically identify and filter out spam and unnatural language from the analyzed reviews, ensuring that users only see trustworthy reviews.

[0926] The server then uses user profile analysis tools to analyze the reviewer's activity. Clustering algorithms (e.g., DBSCAN) are used to detect anomalous behavior, such as reviewers posting a large number of reviews in a short period of time or frequently posting reviews in a particular location. Reviews from reviewers with abnormal activity are also rejected.

[0927] Finally, the server provides filtered and reliable reviews to users through a review providing means. When a user requests a review of a specific store or service from their device, the server can return only reliable reviews. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[0928] As a concrete example, consider a case where a user searches for reviews of "restaurants in Tokyo." The server collects reviews from Google Maps and other review platforms and analyzes them using natural language processing technology. It then filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual behavior. Finally, the filtered, reliable reviews are provided to the user.

[0929] Example prompt sentence:

[0930] "Show me restaurant reviews in Tokyo."

[0931] This process allows users to make decisions based on accurate and reliable information.

[0932] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0933] Step 1:

[0934] The server starts collecting data. It uses the API of the online review platform to obtain review information about the specified store or service. It receives the store or service ID, API key, and authentication information as input, and generates JSON data of the obtained review information as output. This data is then stored in a database.

[0935] Step 2:

[0936] The server applies natural language processing technology. It uses natural language processing libraries (such as NLTK and spaCy) to analyze the collected review information (text data). It takes the review text data as input and outputs the results of spam characteristics, grammatical consistency, and sentiment analysis. Specifically, it performs sentence tokenization, morphological analysis, and sentiment score calculation.

[0937] Step 3:

[0938] The server filters out inappropriate reviews. It uses machine learning algorithms (such as XGBoost) to classify reviews with spam characteristics or unnatural grammar as inappropriate reviews. It takes the results of natural language processing as input and identifies and filters out reviews with scores that are deemed inappropriate. The output is a filtered list of appropriate reviews.

[0939] Step 4:

[0940] The server analyzes the reviewer's user profile, examines the reviewer's posting frequency and activity patterns, and uses a clustering algorithm (such as DBSCAN) to detect fraudulent activity. It takes the reviewer's posting data (posting date and time, location, etc.) as input and filters out reviews from reviewers who are determined to be anomalous. The output is a list of highly reliable reviews that have been further filtered.

[0941] Step 5:

[0942] The server provides the filtered reviews to the user. When a user requests reviews of a specific store or service from their device, the server returns highly reliable reviews. The user inputs the ID of the store or service in question, and the output is highly reliable review information. This allows users to make decisions based on accurate and reliable information.

[0943] Through the above steps, the system can reliably filter review information collected from online review platforms and provide accurate information to users.

[0944] (Application example 1)

[0945] 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."

[0946] Some reviews provided on online review platforms are inappropriate or unreliable, making it difficult for users to make decisions based on accurate information. This problem is particularly pronounced when purchasing products on e-commerce platforms, where users need to make purchasing decisions based on unbiased and reliable reviews.

[0947] 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.

[0948] In this invention, the server includes a data collection means, a natural language processing means, an inappropriate review exclusion means, a user profile analysis means, a reliable review providing means, a product identification means, and a review data display means, thereby providing reliable reviews to users and enabling users to make decisions based on more accurate information.

[0949] "Data Collection Means" means means for collecting review information from online review platforms or e-commerce platforms.

[0950] The "natural language processing means" is a means for analyzing collected reviews and performing spam characteristics, grammatical consistency, sentiment analysis, and context analysis.

[0951] "Inappropriate review filtering" refers to filtering out reviews that have spam characteristics or contain unnatural grammar or extreme emotional expressions.

[0952] "User profile analysis means" is a means of analyzing whether reviewers are engaging in unnatural activities.

[0953] A "means for providing reliable reviews" is a means for providing filtered, reliable reviews to users.

[0954] A "product identification means" is a means that allows a user to search for reviews about a particular product.

[0955] The "review data display means" is a means for visually displaying reliable reviews to the user.

[0956] This invention is a system that reliably filters review information collected from online review platforms or e-commerce platforms and provides users with the most reliable reviews. Implementing this system requires a server, a user terminal, and a program for linking them.

[0957] 1. System Program Overview

[0958] The server executes a program including a data collection means, a natural language processing means, a means for excluding inappropriate reviews, a means for analyzing user profiles, a means for providing reliable reviews, a means for identifying products, and a means for displaying review data. The program performs the following processes:

[0959] 2. Hardware and Software Used

[0960] Hardware:

[0961] Cloud server: The central hardware that collects and analyzes data.

[0962] Smartphone: The device where users search for reviews and view filtered reviews.

[0963] software:

[0964] Python: A programming language for natural language processing and data analysis.

[0965] TextBlob: A natural language processing library for performing sentiment analysis on reviews.

[0966] NLTK: A natural language processing library used to analyze text data.

[0967] Requests: An HTTP library for collecting review information through an API.

[0968] 3. Program processing explanation

[0969] The server uses the data collection means to collect review information from the online review platform or the e-commerce platform via an API, and the collected review information is stored on the server as a data set.

[0970] Next, we analyze the collected reviews using natural language processing tools. Specifically, we use TextBlob to analyze the sentiment and grammatical consistency of each review, and to detect spam characteristics. We also use NLTK for contextual analysis.

[0971] As a means of excluding inappropriate reviews, we remove reviews that have spam characteristics, reviews that lack grammar consistency, and reviews that contain extreme emotional expressions from the list. We also use user profile analysis to detect unusual reviewer activity and exclude such reviews.

[0972] Finally, the reliable review providing means provides filtered reliable reviews to the user's smartphone. When the user searches for a specific product, the product is identified by the product identification means and visually displayed by the review data display means.

[0973] 4. Examples of concrete examples and prompts

[0974] Example of a user purchasing an iPhone 12:

[0975] A user searches for "iPhone 12" reviews on their smartphone.

[0976] A server collects reviews from an online review platform.

[0977] The collected reviews are analyzed using natural language processing technology to determine spam characteristics, grammatical consistency, sentiment analysis, and context analysis.

[0978] Inappropriate or unreliable reviews will be filtered out.

[0979] Appropriate reviews are filtered and presented to users as trusted reviews.

[0980] Example prompt for a generative AI model:

[0981] Product being reviewed: iPhone 12

[0982] Review List:

[0983] 1. Review ID: 001

[0984] It says: "This product is great. I'm happy with the price and quality."

[0985] Sentiment analysis result: Positive polarity (0.8)

[0986] 2. Review ID: 002

[0987] Content: "spam, fake review"

[0988] Sentiment analysis results: Neutral (0)

[0989] 3. Review ID: 003

[0990] What it said: "Very good product, but a little pricey. Service was good."

[0991] Sentiment analysis result: Positive polarity (0.5)

[0992] This way, users can make decisions based on reliable reviews.

[0993] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0994] Step 1:

[0995] Data collection

[0996] The server uses the data collection means to obtain review information from online review platforms or e-commerce platforms via API. The input is the API endpoint of each platform, and the output is a dataset of collected reviews. Specifically, it sends an HTTP request and stores the obtained data in JSON format on the server.

[0997] Step 2:

[0998] Natural Language Processing

[0999] The server uses natural language processing tools to analyze the collected reviews. The input is a dataset of reviews obtained by the data collection tools, and the output is a dataset of analyzed reviews. Specifically, it uses the TextBlob library to analyze the text of each review and perform sentiment analysis (positive, negative, neutral), grammatical consistency checks, and spam detection.

[1000] Step 3:

[1001] Inappropriate review exclusion

[1002] The server uses inappropriate review filtering methods to filter out reviews with spam characteristics, unnatural grammar, or extreme emotional expressions. The input is a dataset of reviews analyzed using natural language processing methods, and the output is a dataset of reliable reviews. Specifically, it filters out reviews that have been detected as spam characteristics, grammatically incorrect reviews, and reviews with extremely biased emotions.

[1003] Step 4:

[1004] User profile analysis

[1005] The server uses user profile analysis to analyze reviewer activity and detect unusual activity. The input is a filtered review dataset, and the output is a more reliable review dataset. Specifically, it detects cases where a large number of reviews are posted by the same user in a short period of time, or where reviews are frequently posted in a specific location, and excludes those reviews.

[1006] Step 5:

[1007] Providing reliable reviews

[1008] The server uses the reliable review providing means to provide filtered reliable reviews to the user. The input is the final filtered review dataset, and the output is the reliable reviews displayed on the user's device. Specifically, based on the user's request, the server retrieves reliable reviews related to a specific product from the database and sends them in JSON format to the user's smartphone.

[1009] Step 6:

[1010] Product Identification

[1011] The user's device uses a product identification means to search for a specific product. The input is the product name and product ID searched by the user, and the output is a list of reviews related to that product. Specifically, when the user enters the product name into the smartphone app, the ID corresponding to that product is sent to the server.

[1012] Step 7:

[1013] Review data display

[1014] The user's device visually displays the reliable reviews using the review data display means. The input is a dataset of reliable reviews obtained from the server, and the output is review information displayed on the user's smartphone screen. Specifically, the application formats the obtained reviews and displays them in a format that is easy for the user to view.

[1015] The above processing steps allow users to select products with confidence based on highly reliable reviews.

[1016] 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.

[1017] This invention combines a system that reliably filters review information collected from online review platforms with an emotion engine that recognizes user emotions. The system operates as follows between a server, a terminal, and a user.

[1018] First, the server collects review information from online review platforms. Specifically, the server uses a specific API to obtain review data from Google Maps and other review platforms. The collected data includes user IDs, review content, rating points, posting date and time, etc.

[1019] The server then analyzes the collected review information using natural language processing technology. Using natural language processing techniques, the review text undergoes spam filtering, grammar consistency checks, and sentiment analysis. This analysis identifies whether each review is spam or contains unnatural content.

[1020] The server then uses a sentiment engine to analyze the text data in the review and identify positive, negative, or neutral sentiment. The sentiment engine evaluates the review's sentiment by analyzing the frequency of specific keywords and phrases, the positive / negative tendencies of the context, and other factors. During this process, if a review is very positive or very negative, it is identified as spam or unnatural.

[1021] Next, the server identifies and filters out inappropriate reviews. Inappropriate review filtering measures remove reviews with spam characteristics, poor grammar, or extreme sentiment from the list. For example, reviews containing extreme sentiment such as "This restaurant is amazing!!!" or "I'll never go there again."

[1022] The server then analyzes the reviewer's user profile, using user profile analysis tools to analyze the reviewer's posting frequency, geographic location, and the recency of the reviewer account, among other things. This analysis detects unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting in multiple locations.

[1023] Finally, the server provides users with filtered, reliable reviews. When a user requests a review of a specific store or service from their device, the server provides only reliable reviews. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[1024] Examples:

[1025] A user uses a device to search for reviews of "restaurants in Tokyo."

[1026] The server collects reviews from Google Maps and other review platforms.

[1027] The collected reviews are analyzed using natural language processing technology and an emotion engine to detect spam characteristics and extreme emotions.

[1028] The server filters out inappropriate reviews and analyzes user profiles to detect unusual activity.

[1029] Only trusted reviews resulting from the filtering are provided to the user.

[1030] This is the implementation form of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[1031] The processing flow will be explained below.

[1032] Step 1:

[1033] The server collects review information from online review platforms. Specifically, it uses the APIs of Google Maps and other review platforms to obtain reviews of specified stores and services. This review information includes user IDs, review content, ratings, posting dates, etc.

[1034] Step 2:

[1035] The server analyzes the collected reviews using natural language processing. Specifically, it uses a natural language processing library to analyze the text data and perform spam characteristics, grammatical consistency, and sentiment analysis. For example, it detects the presence of specific spam keywords and grammatical errors, and determines whether the sentiment of the comment is positive, negative, or neutral.

[1036] Step 3:

[1037] The server uses a sentiment engine to evaluate the detailed sentiment of the analyzed reviews. The sentiment engine analyzes specific keywords and phrases within the review to generate an overall sentiment score. Based on this score, the server identifies whether the review is excessively positive or negative. Based on this result, the server can determine whether an overly emotional review is inappropriate.

[1038] Step 4:

[1039] The server identifies and filters out inappropriate reviews. It uses inappropriate review filtering to remove reviews that have spammy characteristics, poor grammar, or extreme sentiments that are identified by the sentiment engine. For example, reviews like "This store is great!!!" and "I'll never go there again" are filtered out.

[1040] Step 5:

[1041] The server analyzes the reviewer's user profile. Using user profile analysis methods, it checks the reviewer's posting frequency, geographic location, account recency, etc. This profile analysis identifies unusual activity when the same user posts a large number of reviews in a short period of time or frequently posts reviews in different locations.

[1042] Step 6:

[1043] The server provides users with filtered, reliable reviews. When a user requests reviews for a specific store or service using their device, the server returns only the filtered, reliable reviews, allowing users to make decisions based on accurate information.

[1044] Examples:

[1045] A user uses a device to search for reviews of "restaurants in Tokyo."

[1046] The server collects review information from Google Maps and other review platforms.

[1047] The collected reviews are analyzed using natural language processing tools and an emotion engine to detect spam characteristics and emotional extremes.

[1048] The server filters out inappropriate reviews and analyzes user profiles to eliminate reviews from reviewers with unusual behavior.

[1049] The server provides filtered and trusted reviews to the user.

[1050] These are the specific processing steps of this system. This process allows users to make accurate decisions based on reliable reviews, and companies and stores can receive fair evaluations.

[1051] Example 2

[1052] 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."

[1053] Online review platforms are often filled with reviews containing unreliable information, spam reviews, and unnatural emotional expressions. The presence of such inappropriate reviews makes it difficult for users to make accurate decisions. Furthermore, reviewer credibility is not evaluated, and inappropriate activity may be overlooked. Therefore, a system is needed that provides reliable reviews and enables users to make accurate decisions.

[1054] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1055] In this invention, the server includes a data collection means, a natural language processing means, a sentiment analysis means, an inappropriate review excluding means, a user profile analysis means, and a review providing means, thereby enabling the server to reliably filter review information collected from online review platforms and provide accurate and fair reviews to users.

[1056] "Data collection means" refers to a device that has the function of collecting review information from an online review platform.

[1057] The "natural language processing means" is a device that has the function of analyzing collected reviews and performing spam characteristics, grammatical consistency, and sentiment analysis.

[1058] The "sentiment analysis means" is a device that has the function of analyzing text data in reviews and identifying positive, negative, and neutral emotions.

[1059] An "inappropriate review filtering method" is a device that has the function of excluding reviews that have spam characteristics, poor grammar, or extreme emotions from the list.

[1060] A "user profile analysis means" is a device that has the function of analyzing a reviewer's posting frequency, geographic location, the recency of the reviewer account, etc., and detecting any unusual activity.

[1061] The "review providing means" is a device that has the function of providing filtered and reliable reviews to users.

[1062] This invention combines a system for reliably filtering review information collected from online review platforms with an emotion engine that recognizes user emotions. The system operates among a server, a terminal, and a user as follows: The server processes and calculates data using specific hardware and software.

[1063] First, the server collects review information from an online review platform. Specifically, the server obtains review data using a specific API (e.g., online review platform API). The collected data includes user IDs, review content, rating points, posting date and time, etc.

[1064] The server then analyzes the collected review information using natural language processing technology. The natural language processing uses Python natural language processing libraries such as NLTK and spaCy. The server then performs spam filtering, grammar consistency checks, and sentiment analysis. This allows it to determine whether each review is spam or contains unnatural content.

[1065] Additionally, the server uses a sentiment engine (e.g., natural language sentiment API) to analyze the text data in the review and identify positive, negative, or neutral sentiment. The sentiment engine evaluates the sentiment of the review by analyzing the frequency of specific keywords and phrases, the positive / negative tendencies of the context, etc. During this process, if a review is very positive or very negative, it is identified as spam or unnatural.

[1066] Next, the server identifies and filters out inappropriate reviews. Inappropriate review filtering measures remove reviews with spam characteristics, poor grammar, or extreme sentiment from the list. For example, reviews containing extreme sentiment such as "This restaurant is amazing!!!" or "I'll never go there again."

[1067] The server then analyzes the reviewer's user profile. Using user profile analysis methods, the server analyzes the reviewer's posting frequency, geographic location, and the recency of the reviewer account. For example, it detects unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting in multiple locations. This is done using the database management systems MySQL and Apache Cassandra.

[1068] Finally, the server provides the filtered, trusted reviews to the user. When a user requests a review of a specific store or service from their device, the server sends only trusted reviews to the device. This allows users to make decisions based on accurate information, and companies and stores can receive fair evaluations.

[1069] Examples:

[1070] A user uses a device to search for reviews of "restaurants in Tokyo."

[1071] The server collects reviews from an online review platform API.

[1072] The collected reviews are analyzed using natural language processing technology and an emotion engine to detect spam characteristics and extreme emotions.

[1073] The server filters out inappropriate reviews and analyzes user profiles to detect unusual activity.

[1074] Only trusted reviews resulting from the filtering are provided to the user.

[1075] Example prompt sentence:

[1076] "Write a program to identify users who submit a large number of reviews in a certain period of time and detect any unusual activity."

[1077] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1078] Step 1:

[1079] The server uses the API of the online review platform to collect review information. As input, it receives data such as the user ID, review content, rating, and posting date and time. Based on this, it stores the review information in a database. This is how the review information is collected.

[1080] Step 2:

[1081] The server analyzes the collected review information using a Python natural language processing library (e.g., NLTK or spaCy). The review content is given as input. First, the review text is preprocessed (tokenization, stop word removal), then spam filtering (Bayesian filtering, etc.) and grammatical consistency check (generating a parse tree). This allows the spam characteristics and grammatical consistency to be evaluated.

[1082] Step 3:

[1083] The server analyzes the sentiment of the reviews using a sentiment engine (e.g., a natural language sentiment API). The preprocessed review content is given as input. The sentiment engine analyzes the frequency and context of specific keywords and phrases, and outputs a sentiment score as positive, negative, or neutral. This classifies the sentiment of each review.

[1084] Step 4:

[1085] The server identifies and filters out inappropriate reviews. It receives inputs such as sentiment scores, spam characteristics, and grammatical consistency assessment results. The inappropriate review filtering method filters out reviews with spam characteristics or extreme sentiment, eliminating unreliable reviews.

[1086] Step 5:

[1087] The server analyzes the reviewer's user profile. As input, it takes information such as the reviewer's posting frequency, geographic location, and account recency. It uses a database management system (e.g., MySQL or Apache Cassandra) to detect unusual activity, such as users posting a large number of reviews in a short period of time or users posting reviews frequently in different locations. This allows it to assess the reviewer's trustworthiness.

[1088] Step 6:

[1089] A user requests reviews of a specific store or service using a device. A search keyword (e.g., "restaurants in Tokyo") is given as input. The server then provides filtered, trusted reviews, allowing users to make decisions based on accurate information.

[1090] (Application example 2)

[1091] 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."

[1092] Conventional review systems often contain spam reviews or reviews with extreme emotional expressions, making it difficult for users to obtain reliable reviews. Furthermore, the credibility of reviews based on reviewer profiles and activity histories is often insufficient, negatively impacting users' judgments. Furthermore, there has been no consistent method to effectively filter these reviews and provide only reliable reviews.

[1093] 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 a data collection means, a natural language processing means, an inappropriate review exclusion means, a user profile analysis means, a review providing means, a sentiment analysis means, and an application means to be installed on the terminal. This makes it possible to analyze review information collected from an online review platform, exclude spam and inappropriate reviews, and display only reliable reviews on the user's terminal.

[1094] "Data Collection Instruments" means instruments capable of collecting review information from online review platforms.

[1095] "Natural language processing means" refers to means for analyzing collected review information and performing spam characteristics, grammatical consistency, and sentiment analysis.

[1096] "Measures to filter out inappropriate reviews" are measures to detect and filter out reviews that have spam-like characteristics, reviews that contain extreme emotional expressions, and reviews with unnatural grammar.

[1097] "User profile analysis means" refers to means for analyzing profile information such as reviewer posting frequency, geographic location, and account recency.

[1098] The "review providing means" is a means for providing filtered and reliable review information to users.

[1099] The "sentiment analysis means" is a means for analyzing the content of a review and identifying positive, negative, or neutral sentiment.

[1100] The "application means installed on the terminal" refers to a means including an application that provides only reliable reviews when review information is displayed on the terminal.

[1101] This invention relates to a system that reliably filters review information collected from online review platforms and provides it to users. This system mainly consists of three elements: a server, a terminal, and a user.

[1102] The server collects review information from online review platforms. Specifically, the server uses a specific API to obtain review information from various review platforms. The collected data includes user IDs, review content, rating points, posting date and time, etc.

[1103] The server analyzes the collected review information using natural language processing. Utilizing natural language processing technology, the review content is subjected to spam filtering, grammar consistency checks, and sentiment analysis. This analysis determines whether each review is spam or contains unnatural content.

[1104] The server then uses sentiment analysis to analyze the review content and identify positive, negative, or neutral sentiment. Sentiment analysis is performed by analyzing the frequency of specific keywords and phrases, as well as contextual sentiment trends. During this process, if a review is extremely positive or negative, it is identified as spam or unnatural.

[1105] The server then detects and filters out inappropriate reviews, such as spam, reviews with extreme emotional expressions, and reviews with poor grammar, eliminating reviews with low credibility from the list.

[1106] The server also analyzes reviewer profiles using user profile analysis tools, which analyze reviewer posting frequency, geographic location, account recency, etc., to detect unusual activity, such as the same user posting a large number of reviews in a short period of time or frequently posting from multiple different locations.

[1107] The terminal displays reviews of products and services searched for by the user through an application means installed on the terminal. The application has a function of displaying only reliable reviews sent from the server, allowing the user to make decisions based on reliable reviews.

[1108] For example, a user searches for a smartphone case on an online shopping site using a device. The server collects review information and analyzes the reviews using sentiment analysis and natural language processing. Inappropriate reviews are filtered out, and only reliable reviews are provided to the user's device after analyzing the user profile.

[1109] The program to realize this process is implemented using Python, Google Cloud Natural Language API, IBM Watson Tone Analyzer API, etc. This system enables users to make accurate decisions based on reliable reviews, and also enables companies and service providers to receive fair evaluations.

[1110] Example prompt: "Create a Python program that uses the data obtained from the review collection API call to analyze the review content using the Google Cloud Natural Language API and IBM Watson Tone Analyzer API, filter out spam and inappropriate reviews, and provide only trustworthy reviews to users."

[1111] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1112] Step 1:

[1113] Collecting review information

[1114] Subject: Server

[1115] Description: The server uses the API of an online review platform to collect review information for a specific product. For example, to obtain reviews for a specific product, the server calls the API and obtains data including the review content, user ID, rating, posting date and time, etc.

[1116] Input: Product ID given to the API

[1117] Output: A list of retrieved review information

[1118] Step 2:

[1119] Review analysis using natural language processing

[1120] Subject: Server

[1121] Description: The server analyzes the collected review information using a natural language processing engine (e.g., Google Cloud Natural Language API). Specifically, it evaluates the credibility of each review by filtering spam, checking grammar consistency, and analyzing sentiment.

[1122] Input: List of review information

[1123] Output: A list of parsed review information (with spam characteristics, grammar consistency, and sentiment scores)

[1124] Step 3:

[1125] Emotion analysis

[1126] Subject: Server

[1127] Description: The server uses a sentiment analysis engine to classify review content as positive, negative, or neutral. This process involves evaluating the frequency of specific keywords and phrases, the positive / negative tendencies of the context, and other factors to generate a sentiment score. Specifically, it uses the IBM Watson Tone Analyzer API.

[1128] Input: A list of parsed review information (with spam characteristics and grammar consistency)

[1129] Output: A list of reviews with sentiment scores

[1130] Step 4:

[1131] Filtering out inappropriate reviews

[1132] Subject: Server

[1133] Description: The server detects and filters out inappropriate reviews based on the sentiment score obtained by the sentiment analysis method. Reviews that are excessively positive or negative, spam-like, or have poor grammar are removed.

[1134] Input: A list of reviews with sentiment scores

[1135] Output: A list of filtered reviews

[1136] Step 5:

[1137] User profile analysis

[1138] Subject: Server

[1139] Description: The server analyzes the reviewer's user profile, which includes posting frequency, geographic location, account recency, etc. If a reviewer posts a large number of reviews in a short period of time or frequently from multiple different locations, this is detected as unusual activity.

[1140] Input: A filtered list of reviews

[1141] Output: A list of reliable reviews

[1142] Step 6:

[1143] Providing a review

[1144] Subject: Terminal

[1145] Description: Provide users with reliable reviews through an application installed on a terminal, which displays reliable review information received from a server to the user.

[1146] Input: A list of reliable reviews

[1147] Output: Reliable review information provided to users

[1148] In this way, the server and the terminal work together to realize a system that provides highly reliable review information to users.

[1149] 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.

[1150] 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.

[1151] 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.

[1152] 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.

[1153] 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.

[1154] 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.

[1155] 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).

[1156] 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.

[1157] 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."

[1158] 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.

[1159] 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).

[1160] 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.

[1161] 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.

[1162] 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.

[1163] 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.

[1164] 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.

[1165] 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.

[1166] 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.

[1167] 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.

[1168] 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.

[1169] 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.

[1170] The following is further disclosed regarding the above embodiment.

[1171] (Claim 1)

[1172] data collection means;

[1173] natural language processing means;

[1174] Inappropriate review exclusion methods and

[1175] User profile analysis means;

[1176] A means for providing a review;

[1177] A system including:

[1178] (Claim 2)

[1179] 2. The system of claim 1, wherein the data collection means collects review information from an online review platform.

[1180] (Claim 3)

[1181] 2. The system according to claim 1, wherein the natural language processing means analyzes the collected reviews to determine spam characteristics, grammatical consistency, and sentiment analysis.

[1182] "Example 1"

[1183] (Claim 1)

[1184] data collection means;

[1185] natural language processing means;

[1186] Inappropriate review exclusion methods and

[1187] User profile analysis means;

[1188] A means for providing a review;

[1189] A system including:

[1190] (Claim 2)

[1191] 2. The system of claim 1, wherein the data collection means collects review information from an online review platform.

[1192] (Claim 3)

[1193] 2. The system according to claim 1, wherein the natural language processing means analyzes the collected reviews to determine spam characteristics, grammatical consistency, and sentiment analysis.

[1194] (Claim 4)

[1195] 10. The system of claim 1, wherein the inappropriate review filtering means uses a machine learning algorithm to identify and filter out inappropriate reviews.

[1196] (Claim 5)

[1197] 10. The system of claim 1, wherein the user profile analysis means uses a clustering algorithm to detect abnormal reviewer activity.

[1198] (Claim 6)

[1199] 2. The system according to claim 1, wherein the review providing means provides filtered, reliable reviews to the terminal.

[1200] "Application Example 1"

[1201] (Claim 1)

[1202] data collection means;

[1203] natural language processing means;

[1204] Inappropriate review exclusion methods and

[1205] User profile analysis means;

[1206] A reliable way to provide reviews and

[1207] A product identification means;

[1208] review data display means;

[1209] A system including:

[1210] (Claim 2)

[1211] The system of claim 1 , wherein the data collection means collects review information from an online review platform or an e-commerce platform.

[1212] (Claim 3)

[1213] 2. The system of claim 1, wherein the natural language processing means analyzes the collected reviews to determine spam characteristics, grammatical consistency, sentiment analysis, and context analysis.

[1214] "Example 2: Combining Emotion Engines"

[1215] (Claim 1)

[1216] data collection means;

[1217] natural language processing means;

[1218] A sentiment analysis means;

[1219] Inappropriate review exclusion methods and

[1220] User profile analysis means;

[1221] A means for providing a review;

[1222] A system including:

[1223] (Claim 2)

[1224] 2. The system of claim 1, wherein the data collection means collects review information from an online review platform.

[1225] (Claim 3)

[1226] 2. The system according to claim 1, wherein the natural language processing means analyzes the collected reviews to determine spam characteristics, grammatical consistency, and sentiment analysis.

[1227] "Application example 2 when combining emotion engines"

[1228] (Claim 1)

[1229] data collection means;

[1230] natural language processing means;

[1231] Inappropriate review exclusion methods and

[1232] User profile analysis means;

[1233] A means for providing a review;

[1234] A sentiment analysis means;

[1235] application means installed on the terminal;

[1236] A system including:

[1237] (Claim 2)

[1238] 2. The system of claim 1, wherein the data collection means collects review information from an online review platform.

[1239] (Claim 3)

[1240] 2. The system according to claim 1, wherein the natural language processing means analyzes the collected reviews to determine spam characteristics, grammatical consistency, and sentiment analysis.

[1241] (Claim 4)

[1242] 2. The system according to claim 1, wherein the sentiment analysis means analyzes review content and identifies positive, negative, and neutral sentiments.

[1243] (Claim 5)

[1244] 2. The system according to claim 1, wherein the application means installed on the terminal includes an application for displaying only highly reliable reviews when displaying reviews on the terminal. [Explanation of symbols]

[1245] 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. data collection means; natural language processing means; Inappropriate review exclusion methods and User profile analysis means; A means for providing a review; A system including:

2. The system of claim 1 , wherein the data collection means collects review information from an online review platform.

3. The system according to claim 1 , wherein the natural language processing means analyzes the collected reviews to determine spam characteristics, grammatical consistency, and sentiment analysis.

Images