System

Abstract

A system is provided.SOLUTION: A system comprising: means for collecting user data; means for normalizing the collected data; means for training a machine learning algorithm using the normalized data; means for recommending content based on user preferences using the trained model; means for supplementing a recommendation list with word-of-mouth information; and means for displaying the recommendation list on a user terminal.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] In recent years, the diversification and volume of digital content has increased, making it difficult for users to discover content that matches their preferences. Conventional recommendation systems tend to provide mainly popular content and may be unable to suggest niche content that users may not be aware of. Furthermore, many systems do not fully utilize user feedback, making it difficult to provide accurate recommendations that match user preferences. The purpose of this invention is to solve these problems and provide users with the joy and excitement of discovering new content. [Means for solving the problem]

[0005] The present invention provides a system including a means for collecting user data, a means for normalizing the collected data, a means for training a machine learning algorithm using the normalized data, a means for recommending content based on user preferences using the trained model, a means for supplementing a recommendation list with word-of-mouth information, and a means for displaying the recommendation list on a user terminal. Furthermore, by further including a means for collecting user feedback, more accurate recommendations based on user preferences can be realized. Furthermore, by using collaborative filtering in the machine learning algorithm, data from a group of users with similar preferences can be utilized to provide users with recommendations that include a wide range of content, including niche content. This allows users to discover new content that would be difficult to find on their own and enjoy a wider variety of entertainment.

[0006] "User data" is digital information related to a user's behavior and preferences, such as their viewing history, ratings, and search history.

[0007] "Collection means" means software or hardware means for collecting User Data in real time or periodically and storing it in a database.

[0008] A "normalization method" is a process or algorithm used to convert data of different formats or scales into a consistent format.

[0009] "Machine learning algorithm" is a general term for mathematical models and computational methods that automatically learn from data and make predictions and classifications.

[0010] "Training" is the process of using collected data to improve machine learning algorithms and increase their prediction accuracy.

[0011] A "recommendation tool" is a system or algorithm that automatically suggests content based on a user's preferences.

[0012] "Word-of-mouth information" is feedback information about content, such as ratings, comments, and reviews provided by other users.

[0013] "Means for displaying" refers to an interface function for visually presenting content and recommendation lists on the user's device.

[0014] "Collaborative filtering" is an algorithm that learns from other users with similar preferences and makes recommendations based on a user's past behavioral data.

[0015] "Means for collecting feedback" refers to the process of collecting user-provided ratings and comments as data and reflecting them in the system.

[0016] A "similar user group" is a group of users who share common preferences and behavioral patterns based on data such as viewing history and ratings.

[0017] A "recommendation list" is a list of content suggested to a user, and is automatically generated based on the user's preferences and interests. [Brief explanation of the drawings]

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

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

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

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

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

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

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

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

[0026] [First embodiment]

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

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

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

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

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

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

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

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

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

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

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

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

[0039] The present invention is a system for automatically recommending content based on a user's preferences, and an embodiment thereof will be described in detail below.

[0040] User Data Collection

[0041] The means of collection is for the server to collect data such as users' viewing history, ratings, search history, etc. in real time or periodically. This is done based on the content viewed by the user and the ratings given.

[0042] Example: A server automatically records viewing history data and receives a rating score each time a user watches a movie.

[0043] Data normalization

[0044] Normalization is the process of converting collected data into a consistent format, making data of different formats and scales uniform and easier for algorithms to use.

[0045] Example: The server converts data rated by users on different scales (e.g., 10-point or 5-point) into a scale from 0 to 1.

[0046] Training an AI model

[0047] Next, a training means is used to train the machine learning algorithm based on the collected data, thereby learning user preference patterns and improving prediction accuracy.

[0048] Example: A server trains a collaborative filtering algorithm using data from a group of users with similar tastes to build a content recommendation model tailored to each user's preferences.

[0049] Complementing content recommendations and word-of-mouth information

[0050] Using the trained model, the system proposes content based on user preferences through a recommendation tool. It also collects user reviews and adds them to the recommendation list, allowing users to easily refer to other users' opinions and ratings.

[0051] Example: The server recommends two movies, "Action Movie A" and "Drama Movie B," based on the user's preferences, and adds evaluation comments from other users.

[0052] Displaying the recommendation list

[0053] The display method is a software interface that visually displays the recommendation list on the user's device, allowing the user to easily find new content.

[0054] Example: When a user opens a smartphone app, a recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[0055] Gathering feedback

[0056] Finally, we use a means to collect feedback, and store user ratings and comments in a database, which can be reflected in future recommendations to improve accuracy.

[0057] Example: If a user rates a movie they watched and leaves a comment saying "the story was good," the server collects that information and uses it for the next recommendation.

[0058] This allows the system of the present invention to provide a richer content experience to the user.

[0059] The processing flow will be explained below.

[0060] Step 1:

[0061] The server collects real-time user behavioral data such as viewing history, ratings, and search history. Every time a user watches a new movie or adds a rating, the information is stored in a database.

[0062] Specific operation: When a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[0063] Step 2:

[0064] The server normalizes the collected data, which involves converting data measured at different scales into a unified format.

[0065] Specific behavior: The server normalizes the user's rating data from a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0066] Step 3:

[0067] The server uses the normalized data to train machine learning algorithms, which learn user preference patterns and improve prediction accuracy.

[0068] How it works: The server trains a collaborative filtering algorithm using the viewing history and rating data of multiple users with similar tastes.

[0069] Step 4:

[0070] Using the trained model, the server automatically recommends content based on the user's preferences, which can range from general popularity to niche content.

[0071] Specific behavior: The server analyzes that the user likes action movies and adds newly released action movies to the recommendation list.

[0072] Step 5:

[0073] The server collects word-of-mouth information about niche content and adds it to a recommendation list. This word-of-mouth information includes ratings and comments from other users.

[0074] Specific behavior: The server collects reviews of the movie "Drama Movie B" in which other users have rated it as "a moving story" and adds it to the recommendation list.

[0075] Step 6:

[0076] The server transmits the generated recommendation list to the user terminal.

[0077] Specific operation: When a user opens the smartphone app, the server synchronizes with the local database and sends the latest recommendation list.

[0078] Step 7:

[0079] The device displays the received recommendation list to the user, who can then select and view content of interest from the list.

[0080] Specific operation: When a user opens a smartphone app, the device displays recommended movies and music on the home screen. The user can tap to watch.

[0081] Step 8:

[0082] Users can watch the recommended content and then add ratings and comments, which will be reflected in the next recommendations.

[0083] Specific behavior: A user watches the movie "Drama Movie B," rates it highly, and comments that "it had a touching ending." This information will be used to generate the next recommendation.

[0084] Step 9:

[0085] The server collects user feedback and stores it in a database, which is then used to generate recommendations next time.

[0086] What happens: The server stores the ratings and comments users leave for "Drama Movie B" in a database and uses them for the next algorithm training.

[0087] These processing steps enable highly accurate content recommendations based on the user's preferences.

[0088] Example 1

[0089] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0090] Current content recommendation systems often lack sufficient accuracy in recommending content based on user preferences. They also often fail to provide effective recommendations to new users. Furthermore, they lack a mechanism for utilizing word-of-mouth information or efficiently collecting user feedback, which means they are unable to respond to real-time changes in user preferences. This can result in a poor user experience and a deterioration in overall system performance.

[0091] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0092] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing the recommendation list with word-of-mouth information, means for displaying the recommendation list on the user's terminal, means for collecting the user's viewing history from the terminal and transmitting it to the server, means for converting the collected data into a consistent format, and means for collecting feedback data based on user ratings and storing it in a database. This allows for accurate understanding of user preferences and enables effective content recommendations even for new users. Furthermore, utilizing word-of-mouth information increases user satisfaction, and collecting feedback continuously improves the accuracy of the system and the user experience.

[0093] "User data" refers to information that indicates a user's behavior and preferences, such as the user's viewing history, ratings, and search history.

[0094] "Collection Methods" refers to the technologies and processes used to obtain and transmit User Data to a server.

[0095] "Normalization methods" refer to techniques and processes that convert collected data into a consistent format and make it easier for algorithms to use.

[0096] A "machine learning algorithm" refers to a computer program that learns patterns and rules from data to make predictions and classifications.

[0097] "Training" refers to the process of training a machine learning algorithm using collected data.

[0098] "Recommendation method" refers to the technology or process that uses a trained model to recommend content based on user preferences.

[0099] "Word-of-mouth information" refers to evaluation comments and reviews of content left by other users.

[0100] A "recommendation list" refers to a list of content recommended to a user.

[0101] "Means for displaying" refers to the technology or process for visually presenting the recommendation list to the user's terminal.

[0102] "User terminal" refers to the device a user uses to interact with the system, such as a smartphone, PC, or tablet.

[0103] "Viewing history" refers to information about content that a user has viewed in the past.

[0104] "Consistent format" refers to a data format that unifies data of different scales and formats and makes it easy for algorithms to use.

[0105] "Feedback data" refers to the rating scores and comments provided by users after viewing content.

[0106] "Database" refers to a system for systematically storing and managing collected user data and feedback data.

[0107] The present invention is a system that automatically recommends content based on a user's preferences, and details thereof will be described below.

[0108] System Overview

[0109] The system mainly consists of the following elements:

[0110] How user data is collected

[0111] A means of normalizing data

[0112] A means of training machine learning algorithms

[0113] Recommendation methods

[0114] A means to complement word-of-mouth information

[0115] How to display a recommendation list

[0116] A means of gathering feedback

[0117] User Data Collection

[0118] The server collects users' viewing history, ratings, and search history in real time or periodically. The data includes the ID, start time, end time, and rating of the content viewed by the user. This data is stored in the server's database and used for analysis.

[0119] Example: When a user watches the movie "Action Movie A," the viewing history is stored in a database on the server.

[0120] Data normalization

[0121] The server converts the collected data into a consistent format, converting ratings from different scales (e.g., 10-point or 5-point scales) into a range between 0 and 1, and standardizing data in other formats, making the data consistent and suitable for training algorithms.

[0122] Example: Ratings of "4 out of 5" and "8 out of 10" are converted to the format "0.8".

[0123] Training an AI model

[0124] The server uses the normalized data to train machine learning algorithms, using techniques such as collaborative filtering and deep learning to learn user preference patterns, resulting in the creation of highly accurate recommendation models.

[0125] Example: The server uses past viewing and rating data to learn each user's preference patterns.

[0126] Complementing content recommendations and word-of-mouth information

[0127] The server uses the trained model to recommend content based on the user's preferences, and adds word-of-mouth information collected from other users to the recommendation list, making it easier for users to refer to other users' ratings.

[0128] Example: Recommend the movies "Action Movie A" and "Drama Movie B" and add ratings and comments from other users.

[0129] Displaying the recommendation list

[0130] The user device (smartphone, PC, tablet, etc.) visually displays the recommendation list sent from the server. When the user opens the application, they can select and watch content of their interest from the recommendation list displayed on the home screen.

[0131] Example: "Action Movie A" and "Drama Movie B" are displayed on the home screen.

[0132] Gathering feedback

[0133] The server stores user feedback (ratings and comments) in a database, which is used to improve the accuracy of recommendations next time.

[0134] Example: If a user watches the movie "Drama Movie B" and then comments that "the story was good," that information is saved on the server and used for the next recommendation.

[0135] Example prompts to input to the generative AI model

[0136] "Please enter the title and rating of the movie you watched:

[0137] Movie Title: "Action Movie A"

[0138] Review comment: "The action scenes were powerful."

[0139] Use this information to generate an up-to-date list of recommendations for your users."

[0140] This system enables highly accurate and personalized content recommendations to users, improving the user experience.

[0141] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0142] Step 1: Collect user data

[0143] Input: Information about the content viewed by the user (ID, viewing start time, end time, rating score)

[0144] How it works: Every time a user watches content, their device automatically sends viewing data to the server, which then records the data in a database. If the user enters a rating after viewing, the rating data is also sent to the server.

[0145] Output: Viewing history and rating data stored in a database

[0146] Step 2: Normalize the data

[0147] Input: Raw data in the database (viewing history, rating data, search history)

[0148] Specific operation: The server performs normalization processing on the collected data. For example, it standardizes all data rated on different scales (10-point or 5-point) to a scale from 0 to 1.

[0149] Output: Normalized, consistent data

[0150] Step 3: Training the AI ​​model

[0151] Input: Normalized viewing history data and rating data

[0152] Specific operation: The server uses the normalized data to train machine learning algorithms (e.g., collaborative filtering and deep learning), extracting features from the data and analyzing user preference patterns to build a model.

[0153] Output: A trained, highly accurate recommendation model

[0154] Step 4: Recommend content and complement reviews

[0155] Input: Trained AI model and customer reviews

[0156] Specific operation: The server uses the trained model to generate a content list optimized for each user. At the same time, it retrieves other users' evaluation comments from a word-of-mouth information database and adds them to the recommendation list.

[0157] Output: Recommendation list supplemented with reviews

[0158] Step 5: View the recommendation list

[0159] Input: Recommendation list sent from the server

[0160] Specific operation: When a user opens the application on a device (such as a smartphone or PC), a recommendation list sent from the server is displayed. The user can select and watch content that interests them from this list.

[0161] Output: Recommendation list displayed on the user's device

[0162] Step 6: Gather feedback

[0163] Input: Rating score and comments entered by users after viewing the content

[0164] How it works: After watching a video, users can leave a rating or comment, and that information is sent from the device to the server, which then stores this feedback in a database and uses it to train the recommendation model next time.

[0165] Output: User feedback data stored in a database

[0166] Through the above processing steps, the system can provide highly accurate content based on the user's preferences.

[0167] (Application example 1)

[0168] 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."

[0169] In today's world, it is extremely difficult for users to find content that suits their preferences from the vast amount of content available. Furthermore, conventional content recommendation systems are inadequately responsive to the specific devices and platforms of individual users. Furthermore, few systems offer a function that allows users to easily refer to word-of-mouth information, leaving users with a lack of reference information when making their selections.

[0170] 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.

[0171] In this invention, the server includes means for collecting user data, means for normalizing the data, means for training with a machine learning algorithm, means for recommending content based on user preferences using a smartphone, and means for supplementing word-of-mouth information. This allows users to easily find content that matches their preferences on their smartphone, and by referring to word-of-mouth information from other users, they can select content that will provide them with greater satisfaction.

[0172] "User data" refers to information such as a user's viewing history, ratings, and search history, and is data that indicates the user's preferences and behavior.

[0173] "Normalization" is the process of converting collected data into a consistent format and unifying data of different scales and formats.

[0174] A "machine learning algorithm" is a technology that analyzes data, learns user behavior patterns and preferences, and makes predictions and classifications.

[0175] "Recommendation" refers to selecting suitable content based on a user's preferences and presenting it to the user.

[0176] "Word-of-mouth information" refers to information such as ratings and comments provided by other users, and is opinions and impressions that can be used as a reference for making a selection.

[0177] A "user terminal" is an electronic device used by a user to display information and perform operations.

[0178] "Feedback" refers to reaction data such as ratings and comments provided by users after use.

[0179] "Collaborative filtering" is a method for recommending optimal content to individual users by comparing their preferences with data from other users who have similar preferences.

[0180] A "smartphone" is a portable information terminal with multiple functions, including communication, information processing, and application execution.

[0181] A "personalized movie recommendation app" is an application that recommends movies and TV dramas based on the preferences of each individual user.

[0182] The present invention is a personalized recommendation system that enables users to efficiently find content such as movies and dramas on their smartphones. The system of the present invention is configured as follows.

[0183] User Data Collection

[0184] The server collects data such as users' viewing history, ratings, and search history in real time or periodically. This provides basic data for understanding users' preferences. Specifically, each time a user watches a movie or drama, the server records the user's viewing history and rating score. Search history is also collected to understand the genres and titles in which the user is interested.

[0185] Data normalization

[0186] The server normalizes the collected data. User ratings are often rated on different scales (e.g., 10-point scale or 5-point scale). This data is converted to a unified scale (ranging from 0 to 1) to make it easier for machine learning algorithms to use.

[0187] Training an AI model

[0188] Using the collected and normalized data, the server trains a machine learning algorithm. This uses a collaborative filtering algorithm to build a content recommendation model for each user based on data from users with similar tastes. The trained model learns the user's preference patterns and can make more accurate recommendations.

[0189] Complementing content recommendations and word-of-mouth information

[0190] Using the trained model, the server recommends movies and TV shows based on the user's preferences to the smartphone. It also collects word-of-mouth information from other users and adds it to the recommendation list, allowing users to refer to other users' opinions and ratings.

[0191] Displaying the recommendation list

[0192] When a user opens the smartphone app, the recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[0193] Gathering feedback

[0194] When users leave ratings and comments on movies and TV shows they have watched, the server collects that feedback and stores it in a database. This information is then reflected in the next recommendation, further improving the system's prediction accuracy.

[0195] Through the above process, the system can provide the user with a personalized content experience. As a specific example, the following prompt sentence can be used:

[0196] Example prompt sentence:

[0197] Based on user ID 1's recent viewing history, what movies would you recommend to him?

[0198] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0199] Step 1: Collect user data

[0200] The server collects data such as the user's viewing history, rating scores, and search history. This data is collected in real time or periodically. For example, every time a user watches a movie, the viewing history and rating score are sent to the server and recorded in a database. The input is the user's viewing history, rating scores, and search history, and the output is the consistently recorded user data.

[0201] Step 2: Normalize the data

[0202] The server normalizes the collected user data into a consistent format. Specifically, it converts data rated on different scales into a scale from 0 to 1. This process unifies the data and makes it easier to use in machine learning algorithms. The input is the rating data on different scales, and the output is data normalized from 0 to 1.

[0203] Step 3: Training the AI ​​model

[0204] The server uses the normalized data to train a machine learning algorithm, which uses a collaborative filtering algorithm to build a content recommendation model based on data from a group of users with similar preferences. The input is the normalized user data, and the output is the trained recommendation model.

[0205] Step 4: Recommend content and complement reviews

[0206] The server uses the trained model to recommend content based on the user's preferences. In addition, word-of-mouth information collected from other users is added to the recommendation list, allowing users to refer to other users' ratings and comments. The input is the trained recommendation model and word-of-mouth information, and the output is a recommendation list with the word-of-mouth information added.

[0207] Step 5: View the recommendation list

[0208] When a user opens the smartphone app, the server sends a recommendation list to the user's device and displays it on the home screen. The user can select and watch content of interest from this list. The input is a recommendation list with added word-of-mouth information, and the output is a recommendation list displayed on the user's smartphone.

[0209] Step 6: Gather feedback

[0210] When users leave ratings or comments on movies or TV shows they have watched, the server collects that feedback and stores it in a database. The feedback data is reflected in the next recommendation. The input is the user's ratings and comments, and the output is the feedback data stored in the database.

[0211] The above is a specific processing flow in the system of the present invention.

[0212] 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.

[0213] The present invention is a system for automatically recommending content based on a user's preferences and emotions, and an embodiment thereof will be described in detail below.

[0214] User Data Collection

[0215] The means of collection is for the server to collect behavioral data such as user viewing history, ratings, and search history in real time or periodically, based on the content viewed by the user and the ratings given.

[0216] Example: If a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[0217] Data normalization

[0218] Normalization is the process of converting collected data into a consistent format, making data of different formats and scales uniform and easier for algorithms to use.

[0219] Example: The server normalizes the data that the user has rated on a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0220] Training an AI model

[0221] Next, a training means is used to train the machine learning algorithm based on the collected data, thereby learning user preference patterns and improving prediction accuracy.

[0222] Example: A server trains a collaborative filtering algorithm using viewing history and rating data from multiple users with similar tastes.

[0223] Complementing content recommendations and word-of-mouth information

[0224] Using the trained model, the system proposes content based on user preferences through a recommendation tool. It also collects user reviews and adds them to the recommendation list, allowing users to easily refer to other users' opinions and ratings.

[0225] Example: The server recommends two movies, "Action Movie A" and "Drama Movie B," based on the user's preferences, and adds evaluation comments from other users.

[0226] Combining Emotion Engines

[0227] Furthermore, it combines an emotion engine to collect user emotional data, which recognizes emotions from users' facial expressions, voice, and text messages, and optimizes the recommendation list based on that data.

[0228] Example: The server monitors the user's facial expressions and tone of voice while watching the movie "Horror Movie C." If the server determines that the user is feeling scared, it will select a less frightening movie for the next recommendation.

[0229] Displaying the recommendation list

[0230] The display method is a software interface that visually displays the recommendation list on the user's device, allowing the user to easily find new content.

[0231] Example: When a user opens a smartphone app, a recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[0232] Gathering feedback

[0233] Finally, we use a feedback collection method to store user ratings and comments in a database, which can be used to improve the accuracy of future recommendations.

[0234] Example: If a user rates a movie they watched and leaves a comment saying "the story was good," the server collects that information and uses it when generating the next recommendation.

[0235] In this way, the system of the present invention can realize highly accurate and personalized content recommendations based on the user's preferences and emotions, and provide the user with a richer content experience.

[0236] The processing flow will be explained below.

[0237] Step 1:

[0238] The server collects real-time user behavioral data such as viewing history, ratings, and search history. Every time a user watches a new movie or adds a rating, the information is stored in a database.

[0239] Specific operation: When a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[0240] Step 2:

[0241] The server normalizes the collected data, which involves converting data measured at different scales into a unified format.

[0242] Specific behavior: The server normalizes the user's rating data from a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0243] Step 3:

[0244] The server uses the normalized data to train machine learning algorithms, which learn user preference patterns and improve prediction accuracy.

[0245] How it works: The server trains a collaborative filtering algorithm using the viewing history and rating data of multiple users with similar tastes.

[0246] Step 4:

[0247] Using the trained model, the server automatically recommends content based on the user's preferences, which can range from general popularity to niche content.

[0248] Specific behavior: The server analyzes that the user likes action movies and adds newly released action movies to the recommendation list.

[0249] Step 5:

[0250] The server collects word-of-mouth information about niche content and adds it to a recommendation list. This word-of-mouth information includes ratings and comments from other users.

[0251] Specific behavior: The server collects reviews of the movie "Drama Movie B" in which other users have rated it as "a moving story" and adds it to the recommendation list.

[0252] Step 6:

[0253] The server collects the user's emotion data using an emotion engine that recognizes the user's emotions, which includes the process of recognizing emotions from the user's facial expressions, voice, and text messages.

[0254] Specific operation: The server analyzes the user's facial expressions and tone of voice while watching the movie "Horror Movie C," and if it determines that the user is feeling scared, it stores that data.

[0255] Step 7:

[0256] The server optimizes the recommendation list based on the emotional data, adjusting the recommendations of relevant content according to the user's emotional state.

[0257] Specific behavior: The server recognizes that the user is prone to fear and generates a recommendation list that reduces the proportion of horror movies and increases the proportion of comedy movies.

[0258] Step 8:

[0259] The server transmits the generated recommendation list to the user terminal.

[0260] Specific operation: When a user opens the smartphone app, the server synchronizes with the local database and sends the latest recommendation list.

[0261] Step 9:

[0262] The terminal displays the received recommendation list to the user, who can then select and view content of interest from the list.

[0263] Specific operation: When a user opens a smartphone app, the device displays recommended movies and music on the home screen. The user can tap to watch.

[0264] Step 10:

[0265] Users can watch the recommended content and then add ratings and comments, which will be reflected in the next recommendations.

[0266] Specific behavior: A user watches the movie "Drama Movie B," rates it highly, and comments that "it had a touching ending." This information will be used to generate the next recommendation.

[0267] Step 11:

[0268] The server collects user feedback and stores it in a database, which is then used to generate recommendations next time.

[0269] What happens: The server stores the ratings and comments users leave for "Drama Movie B" in a database and uses them for the next algorithm training.

[0270] These processing steps enable highly accurate content recommendations based on the user's preferences and emotions.

[0271] Example 2

[0272] 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."

[0273] Conventional content recommendation systems rely on recommendation functions based on users' viewing history and rating data, but do not take into account the user's emotional state. This makes it difficult to provide optimal content to individual users, and there is a need to improve the user experience.

[0274] 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.

[0275] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing the recommendation list with word-of-mouth information, means for collecting emotional data to optimize the recommendation list, means for displaying the recommendation list on a user terminal, and means for collecting feedback, thereby enabling highly personalized recommendations that take into account not only the preferences but also the emotional state of each individual user.

[0276] "Means for collecting user data" refers to a function for collecting behavioral data such as user viewing history, ratings, and search history in real time or periodically.

[0277] "Means for normalizing collected data" refers to the ability to perform a process to convert data of different formats or scales into a consistent format.

[0278] The "means for training a machine learning algorithm" is a function for training a machine learning algorithm using normalized data to improve the accuracy of the model.

[0279] "Means for recommending content based on user preferences" is a function that uses a trained model to suggest content that best suits the user's preferences.

[0280] "Means for supplementing recommendation lists with word-of-mouth information" refers to a function that collects ratings and comments from other users and adds them to the recommendation list.

[0281] "Means of collecting emotional data to optimize recommendation lists" refers to a function that recognizes emotions from the user's facial expressions, voice, and text messages, and uses that data to adjust the recommendation list to be more optimal for the individual.

[0282] The "means for displaying the recommendation list on the user terminal" is a function for visually displaying the recommendation list sent from the server on the interface of the user terminal.

[0283] "Means for collecting feedback" is a function that collects ratings and comments from users, stores that information in a database, and reflects it in the next recommendation.

[0284] "Collaborative filtering" is a machine learning algorithm that recommends appropriate content to individual users based on the viewing history and rating data of multiple users with similar tastes.

[0285] This invention is a system that automatically recommends content based on user preferences and emotions. This system mainly consists of a server and a user terminal. The server is responsible for data collection, data normalization, machine learning algorithm training, emotion data collection, recommendation generation, and feedback collection. The user terminal provides an interface with the user and has the function of displaying a recommendation list.

[0286] User Data Collection

[0287] The server collects user behavioral data such as viewing history, ratings, and search history in real time or periodically. This allows the server to accurately grasp user preferences. For example, if a user watches "Action Movie A" and rates it 4 out of 5, the viewing history and rating are added to the database.

[0288] Data normalization

[0289] The server then processes the collected data to convert it into a consistent format, unifying data from different formats and scales, making it easier for machine learning algorithms to use. For example, data rated by a user on a 10-point scale is normalized to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0290] Training an AI model

[0291] The server trains machine learning algorithms based on the normalized data, which allows it to learn user preference patterns and improve prediction accuracy. For example, a collaborative filtering algorithm is used to learn the viewing history and rating data of multiple users with similar preferences.

[0292] Complementing content recommendations and word-of-mouth information

[0293] Using the trained model, the server recommends content based on the user's preferences. It also supplements the recommendation list with word-of-mouth information to provide other users' opinions and ratings. For example, it recommends "Action Movie A" and "Drama Movie B" to the user, along with other users' ratings and comments.

[0294] Combining Emotion Engines

[0295] The server uses an emotion engine to collect emotional data from the user's facial expressions, voice, and text messages, and optimizes the recommendation list. This process allows the server to provide content that best suits the user's current emotional state. For example, if the user expresses fear while watching "Horror Movie C," the server will select a less frightening movie for the next recommendation list.

[0296] Displaying the recommendation list

[0297] The user device uses a software interface that visually displays the recommendation list sent from the server, allowing the user to easily find new content. For example, when a user opens a smartphone app, the recommendation list is displayed on the home screen.

[0298] Gathering feedback

[0299] The server collects user ratings and comments in a database and reflects this information in the next recommendation. For example, if a user rates a movie they watched and leaves a comment saying "the story was good," that information will be used when generating the next recommendation.

[0300] In this way, the system of the present invention can realize highly accurate and personalized content recommendations based on the user's preferences and emotions, providing the user with a richer content experience.

[0301] Prompt Sentence Examples

[0302] "A user watched the movie 'Action Movie A' and rated it 4 out of 5. Please recommend new action movies based on their viewing history and ratings."

[0303] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0304] Step 1: Collect user data

[0305] The server collects user behavioral data, such as viewing history, ratings, and search history, in real time or periodically. Specifically, when a user watches a particular movie and enters a rating, the information is sent to the server. The input data is the movie title "Action Movie A" and the rating "4." The output is that these data are stored in a database.

[0306] Step 2: Normalize the data

[0307] The server then processes the collected raw user data to convert it into a consistent format. For example, it normalizes the data that the user has rated on a 10-point scale to a 0-1 scale. This unifies data from different formats and scales. The input data is a rating of 4, which is normalized to 0.8, and the normalized data is stored as the output.

[0308] Step 3: Training the AI ​​model

[0309] The server uses the normalized data to train a machine learning algorithm. Specifically, it uses a collaborative filtering algorithm to train the model on the viewing history and rating data of multiple users with similar preferences. The normalized rating data (0.8) and other users' data are used as input, and the output is the trained model.

[0310] Step 4: Recommend content and complement reviews

[0311] Using the trained model, the server recommends content based on the user's preferences. The server also supplements the recommendation list with reviews collected from other users. The input data is the trained model and review data, and the output is a recommendation list tailored to the user. Specifically, "Action Movie A" and "Drama Movie B" are included in the list, and reviews for each are displayed.

[0312] Step 5: Combining Emotion Engines

[0313] The server uses an emotion engine to collect user emotional data and optimize the recommendation list. It recognizes emotions from the user's facial expressions, voice, and text messages and makes adjustments based on that data. The input data is the user's emotional data while watching, and the output is a recommendation list that takes emotions into consideration. Specifically, if the server recognizes that the user is feeling scared, the next movie recommendation will be a less scary one.

[0314] Step 6: Displaying the recommendation list

[0315] The device uses a software interface that visually presents the recommendation list sent from the server. The input is the recommendation list data from the server, and the output is the list displayed on the device screen. When a user opens the smartphone app, a list of related recommended movies and TV shows is displayed on the home screen.

[0316] Step 7: Gather feedback

[0317] The server collects ratings and comments from users and reflects this information in the next recommendation. The input data are user ratings and comments, which are saved in a database and used to generate the next recommendation. The output is an updated database. For example, if a user rates a story as "good," that information is reflected in the server.

[0318] (Application example 2)

[0319] 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."

[0320] Conventional content recommendation systems make recommendations based on user preferences and behavioral data, but do not take into account the user's real-time emotions when recommending content. As a result, content that does not match the user's momentary emotions or state may be recommended, resulting in a poor user experience. The present invention aims to provide a system that combines the user's real-time emotional information to realize highly accurate and personalized content recommendations.

[0321] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0322] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing a recommendation list with word-of-mouth information, means for displaying the recommendation list on a user terminal, means for acquiring emotion data using the user's facial expressions and voice, and means for adjusting the recommendation list in real time based on the emotion data. This makes it possible to recommend more appropriate content based on the user's real-time emotions.

[0323] "User data" refers to information about a user's behavior and preferences, such as their viewing history, ratings, and search history.

[0324] "Means for collecting" refers to a method or device by which the server collects behavioral data such as user viewing history and ratings in real time or periodically.

[0325] "Normalization means" refers to a method or device that converts collected data into a consistent format and unifies data of different formats or scales.

[0326] "Means for training" refers to a method or apparatus that uses normalized data to train a machine learning algorithm.

[0327] "Means for recommending" refers to a method or apparatus that uses a trained model to suggest content based on a user's preferences.

[0328] The "means for supplementing word-of-mouth information" refers to a method or device for collecting opinions and ratings from other users and adding them to a recommendation list.

[0329] The "means for displaying" refers to a software interface that visually presents the recommendation list on the user terminal.

[0330] "Means for acquiring emotion data using facial expressions and voice" refers to a method or device for capturing a user's facial expressions and voice in real time and recognizing their emotions.

[0331] "Means for adjusting a recommendation list in real time" refers to a method or apparatus for optimizing a recommendation list in real time based on acquired emotion data.

[0332] "Machine learning algorithm" refers to an algorithm that learns user preference patterns based on collected data and improves prediction accuracy.

[0333] "Collaborative filtering" refers to a technique for improving recommendation accuracy by using data from multiple users with similar preferences.

[0334] The present invention is a system for automatically recommending content based on a user's preferences and emotions, and an embodiment thereof will be described in detail below.

[0335] System configuration

[0336] The system includes means for collecting user data, normalizing the data, training machine learning algorithms, recommending content, complementing reviews, displaying recommendation lists, acquiring sentiment data, and adjusting recommendations in real time.

[0337] Data collection

[0338] The server collects user behavioral data, such as viewing history, ratings, and search history, in real time or periodically. To do so, it obtains the user's viewing history and ratings from a database. It also collects facial expressions and voice data using the smartphone's camera and microphone.

[0339] Data normalization

[0340] The server converts the collected data into a consistent format, specifically normalizing the rating data to a scale of 0 to 1. This unifies the variations in data format and scale, making it easier for algorithms to use.

[0341] Training machine learning algorithms

[0342] The server uses the normalized data to train a machine learning algorithm, using a collaborative filtering algorithm to build a model using data from multiple users with similar preferences.

[0343] Content recommendations

[0344] The trained model is used to predict user preference patterns and recommend the most suitable content, thereby providing personalized content to users.

[0345] Complementing reviews

[0346] The server collects word-of-mouth information from other users and adds it to the recommendation list, making it easier for users to refer to the opinions and ratings of other users.

[0347] Displaying the recommendation list

[0348] The recommendation list sent from the server is visually displayed on the user's device, allowing the user to select and view content of interest from the list.

[0349] Acquiring emotion data

[0350] The server uses the smartphone's camera and microphone to obtain emotion data from the user's facial expressions and voice, using a neural network model for emotion analysis.

[0351] Real-time recommendation adjustment

[0352] The server adjusts the recommendation list in real time based on the acquired emotional data. For example, if a user is watching a scary movie, the server will recognize that emotional data in real time and recommend less scary content next.

[0353] Specific examples

[0354] If a user feels scared while watching "Horror Movie C," the server will recognize that emotion in real time and recommend a less scary movie for the next viewing, taking into account word-of-mouth reviews from other users to suggest more appropriate content.

[0355] Prompt Sentence Examples

[0356] "If it is detected in the user's viewing history that they felt scared while watching 'Horror Movie C,' recommend movies that are less scary based on their viewing history and rating data. Also, take into account word-of-mouth information from other users."

[0357] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0358] Step 1:

[0359] User Data Collection

[0360] Input: User's viewing history, ratings, search history, facial expression and voice data from the smartphone camera and microphone

[0361] The server collects user behavior data from a database in real time or periodically, and also collects facial expression and voice data using the smartphone's camera and microphone.

[0362] Output: Collected user viewing history, rating data, search history, facial expression data, and voice data

[0363] Step 2:

[0364] Data normalization

[0365] Input: Collected viewing history, rating data, search history

[0366] The server normalizes the collected rating data to a scale of 0 to 1. Specifically, if the rating is out of 10, it converts it to a scale of 0 to 1.

[0367] Output: Normalized rating data

[0368] Step 3:

[0369] Training a machine learning model

[0370] Input: Normalized rating data, viewing history, search history

[0371] The server uses the normalized data to train a collaborative filtering algorithm, building a model using data from multiple users with similar preferences.

[0372] Output: A trained machine learning model

[0373] Step 4:

[0374] Content recommendations

[0375] Input: Trained machine learning model, current user preference data

[0376] The server uses the trained model to recommend optimal content based on the current user preferences.

[0377] Output: Recommendation list

[0378] Step 5:

[0379] Complementing reviews

[0380] Input: Recommendation list, word-of-mouth data from other users

[0381] The server adds word-of-mouth information from other users to the recommendation list, thereby providing other users' opinions and ratings about the recommended content.

[0382] Output: Recommendation list with reviews added

[0383] Step 6:

[0384] Displaying the recommendation list

[0385] Input: Recommendation list with reviews

[0386] The server visually provides a recommendation list to the user's device, allowing the user to select and view content of interest from the list.

[0387] Output: Recommendation list displayed on the user's device

[0388] Step 7:

[0389] Acquiring emotion data

[0390] Input: User's facial expression data, voice data

[0391] The server uses the smartphone's camera and microphone to analyze data obtained from the user's facial expressions and voice using an emotion recognition algorithm, and uses a neural network to estimate the user's emotions.

[0392] Output: Real-time emotion data

[0393] Step 8:

[0394] Real-time recommendation adjustment

[0395] Input: Real-time emotion data, recommendation list

[0396] The server adjusts the recommendation list based on real-time emotion data. For example, if a user feels scared, the server will reflect that emotion and recommend a less scary movie as the next content to watch.

[0397] Output: A refined list of recommendations

[0398] 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.

[0399] 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.

[0400] 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.

[0401] [Second embodiment]

[0402] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0403] 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.

[0404] 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).

[0405] 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.

[0406] 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.

[0407] 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).

[0408] 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. 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.

[0409] 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.

[0410] 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.

[0411] 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.

[0412] In the smart glasses 214, 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.

[0413] 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."

[0414] The present invention is a system for automatically recommending content based on a user's preferences, and an embodiment thereof will be described in detail below.

[0415] User Data Collection

[0416] The means of collection is for the server to collect data such as users' viewing history, ratings, search history, etc. in real time or periodically. This is done based on the content viewed by the user and the ratings given.

[0417] Example: A server automatically records viewing history data and receives a rating score each time a user watches a movie.

[0418] Data normalization

[0419] Normalization is the process of converting collected data into a consistent format, making data of different formats and scales uniform and easier for algorithms to use.

[0420] Example: The server converts data rated by users on different scales (e.g., 10-point or 5-point) into a scale from 0 to 1.

[0421] Training an AI model

[0422] Next, a training means is used to train the machine learning algorithm based on the collected data, thereby learning user preference patterns and improving prediction accuracy.

[0423] Example: A server trains a collaborative filtering algorithm using data from a group of users with similar tastes to build a content recommendation model tailored to each user's preferences.

[0424] Complementing content recommendations and word-of-mouth information

[0425] Using the trained model, the system proposes content based on user preferences through a recommendation tool. It also collects user reviews and adds them to the recommendation list, allowing users to easily refer to other users' opinions and ratings.

[0426] Example: The server recommends two movies, "Action Movie A" and "Drama Movie B," based on the user's preferences, and adds evaluation comments from other users.

[0427] Displaying the recommendation list

[0428] The display method is a software interface that visually displays the recommendation list on the user's device, allowing the user to easily find new content.

[0429] Example: When a user opens a smartphone app, a recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[0430] Gathering feedback

[0431] Finally, we use a means to collect feedback, and store user ratings and comments in a database, which can be reflected in future recommendations to improve accuracy.

[0432] Example: If a user rates a movie they watched and leaves a comment saying "the story was good," the server collects that information and uses it for the next recommendation.

[0433] This allows the system of the present invention to provide a richer content experience to the user.

[0434] The processing flow will be explained below.

[0435] Step 1:

[0436] The server collects real-time user behavioral data such as viewing history, ratings, and search history. Every time a user watches a new movie or adds a rating, the information is stored in a database.

[0437] Specific operation: When a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[0438] Step 2:

[0439] The server normalizes the collected data, which involves converting data measured at different scales into a unified format.

[0440] Specific behavior: The server normalizes the user's rating data from a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0441] Step 3:

[0442] The server uses the normalized data to train machine learning algorithms, which learn user preference patterns and improve prediction accuracy.

[0443] How it works: The server trains a collaborative filtering algorithm using the viewing history and rating data of multiple users with similar tastes.

[0444] Step 4:

[0445] Using the trained model, the server automatically recommends content based on the user's preferences, which can range from general popularity to niche content.

[0446] Specific behavior: The server analyzes that the user likes action movies and adds newly released action movies to the recommendation list.

[0447] Step 5:

[0448] The server collects word-of-mouth information about niche content and adds it to a recommendation list. This word-of-mouth information includes ratings and comments from other users.

[0449] Specific behavior: The server collects reviews of the movie "Drama Movie B" in which other users have rated it as "a moving story" and adds it to the recommendation list.

[0450] Step 6:

[0451] The server transmits the generated recommendation list to the user terminal.

[0452] Specific operation: When a user opens the smartphone app, the server synchronizes with the local database and sends the latest recommendation list.

[0453] Step 7:

[0454] The device displays the received recommendation list to the user, who can then select and view content of interest from the list.

[0455] Specific operation: When a user opens a smartphone app, the device displays recommended movies and music on the home screen. The user can tap to watch.

[0456] Step 8:

[0457] Users can watch the recommended content and then add ratings and comments, which will be reflected in the next recommendations.

[0458] Specific behavior: A user watches the movie "Drama Movie B," rates it highly, and comments that "it had a touching ending." This information will be used to generate the next recommendation.

[0459] Step 9:

[0460] The server collects user feedback and stores it in a database, which is then used to generate recommendations next time.

[0461] What happens: The server stores the ratings and comments users leave for "Drama Movie B" in a database and uses them for the next algorithm training.

[0462] These processing steps enable highly accurate content recommendations based on the user's preferences.

[0463] Example 1

[0464] 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."

[0465] Current content recommendation systems often lack sufficient accuracy in recommending content based on user preferences. They also often fail to provide effective recommendations to new users. Furthermore, they lack a mechanism for utilizing word-of-mouth information or efficiently collecting user feedback, which means they are unable to respond to real-time changes in user preferences. This can result in a poor user experience and a deterioration in overall system performance.

[0466] 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.

[0467] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing the recommendation list with word-of-mouth information, means for displaying the recommendation list on the user's terminal, means for collecting the user's viewing history from the terminal and transmitting it to the server, means for converting the collected data into a consistent format, and means for collecting feedback data based on user ratings and storing it in a database. This allows for accurate understanding of user preferences and enables effective content recommendations even for new users. Furthermore, utilizing word-of-mouth information increases user satisfaction, and collecting feedback continuously improves the accuracy of the system and the user experience.

[0468] "User data" refers to information that indicates a user's behavior and preferences, such as the user's viewing history, ratings, and search history.

[0469] "Collection Methods" refers to the technologies and processes used to obtain and transmit User Data to a server.

[0470] "Normalization methods" refer to techniques and processes that convert collected data into a consistent format and make it easier for algorithms to use.

[0471] A "machine learning algorithm" refers to a computer program that learns patterns and rules from data to make predictions and classifications.

[0472] "Training" refers to the process of training a machine learning algorithm using collected data.

[0473] "Recommendation method" refers to the technology or process that uses a trained model to recommend content based on user preferences.

[0474] "Word-of-mouth information" refers to evaluation comments and reviews of content left by other users.

[0475] A "recommendation list" refers to a list of content recommended to a user.

[0476] "Means for displaying" refers to the technology or process for visually presenting the recommendation list to the user's terminal.

[0477] "User terminal" refers to the device a user uses to interact with the system, such as a smartphone, PC, or tablet.

[0478] "Viewing history" refers to information about content that a user has viewed in the past.

[0479] "Consistent format" refers to a data format that unifies data of different scales and formats and makes it easy for algorithms to use.

[0480] "Feedback data" refers to the rating scores and comments provided by users after viewing content.

[0481] "Database" refers to a system for systematically storing and managing collected user data and feedback data.

[0482] The present invention is a system that automatically recommends content based on a user's preferences, and details thereof will be described below.

[0483] System Overview

[0484] The system mainly consists of the following elements:

[0485] How user data is collected

[0486] A means of normalizing data

[0487] A means of training machine learning algorithms

[0488] Recommendation methods

[0489] A means to complement word-of-mouth information

[0490] How to display a recommendation list

[0491] A means of gathering feedback

[0492] User Data Collection

[0493] The server collects users' viewing history, ratings, and search history in real time or periodically. The data includes the ID, start time, end time, and rating of the content viewed by the user. This data is stored in the server's database and used for analysis.

[0494] Example: When a user watches the movie "Action Movie A," the viewing history is stored in a database on the server.

[0495] Data normalization

[0496] The server converts the collected data into a consistent format, converting ratings from different scales (e.g., 10-point or 5-point scales) into a range between 0 and 1, and standardizing data in other formats, making the data consistent and suitable for training algorithms.

[0497] Example: Ratings of "4 out of 5" and "8 out of 10" are converted to the format "0.8".

[0498] Training an AI model

[0499] The server uses the normalized data to train machine learning algorithms, using techniques such as collaborative filtering and deep learning to learn user preference patterns, resulting in the creation of highly accurate recommendation models.

[0500] Example: The server uses past viewing and rating data to learn each user's preference patterns.

[0501] Complementing content recommendations and word-of-mouth information

[0502] The server uses the trained model to recommend content based on the user's preferences, and adds word-of-mouth information collected from other users to the recommendation list, making it easier for users to refer to other users' ratings.

[0503] Example: Recommend the movies "Action Movie A" and "Drama Movie B" and add ratings and comments from other users.

[0504] Displaying the recommendation list

[0505] The user device (smartphone, PC, tablet, etc.) visually displays the recommendation list sent from the server. When the user opens the application, they can select and watch content of their interest from the recommendation list displayed on the home screen.

[0506] Example: "Action Movie A" and "Drama Movie B" are displayed on the home screen.

[0507] Gathering feedback

[0508] The server stores user feedback (ratings and comments) in a database, which is used to improve the accuracy of recommendations next time.

[0509] Example: If a user watches the movie "Drama Movie B" and then comments that "the story was good," that information is saved on the server and used for the next recommendation.

[0510] Example prompts to input to the generative AI model

[0511] "Please enter the title and rating of the movie you watched:

[0512] Movie Title: "Action Movie A"

[0513] Review comment: "The action scenes were powerful."

[0514] Use this information to generate an up-to-date list of recommendations for your users."

[0515] This system enables highly accurate and personalized content recommendations to users, improving the user experience.

[0516] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0517] Step 1: Collect user data

[0518] Input: Information about the content viewed by the user (ID, viewing start time, end time, rating score)

[0519] How it works: Every time a user watches content, their device automatically sends viewing data to the server, which then records the data in a database. If the user enters a rating after viewing, the rating data is also sent to the server.

[0520] Output: Viewing history and rating data stored in a database

[0521] Step 2: Normalize the data

[0522] Input: Raw data in the database (viewing history, rating data, search history)

[0523] Specific operation: The server performs normalization processing on the collected data. For example, it standardizes all data rated on different scales (10-point or 5-point) to a scale from 0 to 1.

[0524] Output: Normalized, consistent data

[0525] Step 3: Training the AI ​​model

[0526] Input: Normalized viewing history data and rating data

[0527] Specific operation: The server uses the normalized data to train machine learning algorithms (e.g., collaborative filtering and deep learning), extracting features from the data and analyzing user preference patterns to build a model.

[0528] Output: A trained, highly accurate recommendation model

[0529] Step 4: Recommend content and complement reviews

[0530] Input: Trained AI model and customer reviews

[0531] Specific operation: The server uses the trained model to generate a content list optimized for each user. At the same time, it retrieves other users' evaluation comments from a word-of-mouth information database and adds them to the recommendation list.

[0532] Output: Recommendation list supplemented with reviews

[0533] Step 5: View the recommendation list

[0534] Input: Recommendation list sent from the server

[0535] Specific operation: When a user opens the application on a device (such as a smartphone or PC), a recommendation list sent from the server is displayed. The user can select and watch content that interests them from this list.

[0536] Output: Recommendation list displayed on the user's device

[0537] Step 6: Gather feedback

[0538] Input: Rating score and comments entered by users after viewing the content

[0539] How it works: After watching a video, users can leave a rating or comment, and that information is sent from the device to the server, which then stores this feedback in a database and uses it to train the recommendation model next time.

[0540] Output: User feedback data stored in a database

[0541] Through the above processing steps, the system can provide highly accurate content based on the user's preferences.

[0542] (Application example 1)

[0543] 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."

[0544] In today's world, it is extremely difficult for users to find content that suits their preferences from the vast amount of content available. Furthermore, conventional content recommendation systems are inadequately responsive to the specific devices and platforms of individual users. Furthermore, few systems offer a function that allows users to easily refer to word-of-mouth information, leaving users with a lack of reference information when making their selections.

[0545] 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.

[0546] In this invention, the server includes means for collecting user data, means for normalizing the data, means for training with a machine learning algorithm, means for recommending content based on user preferences using a smartphone, and means for supplementing word-of-mouth information. This allows users to easily find content that matches their preferences on their smartphone, and by referring to word-of-mouth information from other users, they can select content that will provide them with greater satisfaction.

[0547] "User data" refers to information such as a user's viewing history, ratings, and search history, and is data that indicates the user's preferences and behavior.

[0548] "Normalization" is the process of converting collected data into a consistent format and unifying data of different scales and formats.

[0549] A "machine learning algorithm" is a technology that analyzes data, learns user behavior patterns and preferences, and makes predictions and classifications.

[0550] "Recommendation" refers to selecting suitable content based on a user's preferences and presenting it to the user.

[0551] "Word-of-mouth information" refers to information such as ratings and comments provided by other users, and is opinions and impressions that can be used as a reference for making a selection.

[0552] A "user terminal" is an electronic device used by a user to display information and perform operations.

[0553] "Feedback" refers to reaction data such as ratings and comments provided by users after use.

[0554] "Collaborative filtering" is a method for recommending optimal content to individual users by comparing their preferences with data from other users who have similar preferences.

[0555] A "smartphone" is a portable information terminal with multiple functions, including communication, information processing, and application execution.

[0556] A "personalized movie recommendation app" is an application that recommends movies and TV dramas based on the preferences of each individual user.

[0557] The present invention is a personalized recommendation system that enables users to efficiently find content such as movies and dramas on their smartphones. The system of the present invention is configured as follows.

[0558] User Data Collection

[0559] The server collects data such as users' viewing history, ratings, and search history in real time or periodically. This provides basic data for understanding users' preferences. Specifically, each time a user watches a movie or drama, the server records the user's viewing history and rating score. Search history is also collected to understand the genres and titles in which the user is interested.

[0560] Data normalization

[0561] The server normalizes the collected data. User ratings are often rated on different scales (e.g., 10-point scale or 5-point scale). This data is converted to a unified scale (ranging from 0 to 1) to make it easier for machine learning algorithms to use.

[0562] Training an AI model

[0563] Using the collected and normalized data, the server trains a machine learning algorithm. This uses a collaborative filtering algorithm to build a content recommendation model for each user based on data from users with similar tastes. The trained model learns the user's preference patterns and can make more accurate recommendations.

[0564] Complementing content recommendations and word-of-mouth information

[0565] Using the trained model, the server recommends movies and TV shows based on the user's preferences to the smartphone. It also collects word-of-mouth information from other users and adds it to the recommendation list, allowing users to refer to other users' opinions and ratings.

[0566] Displaying the recommendation list

[0567] When a user opens the smartphone app, the recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[0568] Gathering feedback

[0569] When users leave ratings and comments on movies and TV shows they have watched, the server collects that feedback and stores it in a database. This information is then reflected in the next recommendation, further improving the system's prediction accuracy.

[0570] Through the above process, the system can provide the user with a personalized content experience. As a specific example, the following prompt sentence can be used:

[0571] Example prompt sentence:

[0572] Based on user ID 1's recent viewing history, what movies would you recommend to him?

[0573] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0574] Step 1: Collect user data

[0575] The server collects data such as the user's viewing history, rating scores, and search history. This data is collected in real time or periodically. For example, every time a user watches a movie, the viewing history and rating score are sent to the server and recorded in a database. The input is the user's viewing history, rating scores, and search history, and the output is the consistently recorded user data.

[0576] Step 2: Normalize the data

[0577] The server normalizes the collected user data into a consistent format. Specifically, it converts data rated on different scales into a scale from 0 to 1. This process unifies the data and makes it easier to use in machine learning algorithms. The input is the rating data on different scales, and the output is data normalized from 0 to 1.

[0578] Step 3: Training the AI ​​model

[0579] The server uses the normalized data to train a machine learning algorithm, which uses a collaborative filtering algorithm to build a content recommendation model based on data from a group of users with similar preferences. The input is the normalized user data, and the output is the trained recommendation model.

[0580] Step 4: Recommend content and complement reviews

[0581] The server uses the trained model to recommend content based on the user's preferences. In addition, word-of-mouth information collected from other users is added to the recommendation list, allowing users to refer to other users' ratings and comments. The input is the trained recommendation model and word-of-mouth information, and the output is a recommendation list with the word-of-mouth information added.

[0582] Step 5: View the recommendation list

[0583] When a user opens the smartphone app, the server sends a recommendation list to the user's device and displays it on the home screen. The user can select and watch content of interest from this list. The input is a recommendation list with added word-of-mouth information, and the output is a recommendation list displayed on the user's smartphone.

[0584] Step 6: Gather feedback

[0585] When users leave ratings or comments on movies or TV shows they have watched, the server collects that feedback and stores it in a database. The feedback data is reflected in the next recommendation. The input is the user's ratings and comments, and the output is the feedback data stored in the database.

[0586] The above is a specific processing flow in the system of the present invention.

[0587] 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.

[0588] The present invention is a system for automatically recommending content based on a user's preferences and emotions, and an embodiment thereof will be described in detail below.

[0589] User Data Collection

[0590] The means of collection is for the server to collect behavioral data such as user viewing history, ratings, and search history in real time or periodically, based on the content viewed by the user and the ratings given.

[0591] Example: If a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[0592] Data normalization

[0593] Normalization is the process of converting collected data into a consistent format, making data of different formats and scales uniform and easier for algorithms to use.

[0594] Example: The server normalizes the data that the user has rated on a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0595] Training an AI model

[0596] Next, a training means is used to train the machine learning algorithm based on the collected data, thereby learning user preference patterns and improving prediction accuracy.

[0597] Example: A server trains a collaborative filtering algorithm using viewing history and rating data from multiple users with similar tastes.

[0598] Complementing content recommendations and word-of-mouth information

[0599] Using the trained model, the system proposes content based on user preferences through a recommendation tool. It also collects user reviews and adds them to the recommendation list, allowing users to easily refer to other users' opinions and ratings.

[0600] Example: The server recommends two movies, "Action Movie A" and "Drama Movie B," based on the user's preferences, and adds evaluation comments from other users.

[0601] Combining Emotion Engines

[0602] Furthermore, it combines an emotion engine to collect user emotional data, which recognizes emotions from users' facial expressions, voice, and text messages, and optimizes the recommendation list based on that data.

[0603] Example: The server monitors the user's facial expressions and tone of voice while watching the movie "Horror Movie C." If the server determines that the user is feeling scared, it will select a less frightening movie for the next recommendation.

[0604] Displaying the recommendation list

[0605] The display method is a software interface that visually displays the recommendation list on the user's device, allowing the user to easily find new content.

[0606] Example: When a user opens a smartphone app, a recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[0607] Gathering feedback

[0608] Finally, we use a feedback collection method to store user ratings and comments in a database, which can be used to improve the accuracy of future recommendations.

[0609] Example: If a user rates a movie they watched and leaves a comment saying "the story was good," the server collects that information and uses it when generating the next recommendation.

[0610] In this way, the system of the present invention can realize highly accurate and personalized content recommendations based on the user's preferences and emotions, and provide the user with a richer content experience.

[0611] The processing flow will be explained below.

[0612] Step 1:

[0613] The server collects real-time user behavioral data such as viewing history, ratings, and search history. Every time a user watches a new movie or adds a rating, the information is stored in a database.

[0614] Specific operation: When a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[0615] Step 2:

[0616] The server normalizes the collected data, which involves converting data measured at different scales into a unified format.

[0617] Specific behavior: The server normalizes the user's rating data from a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0618] Step 3:

[0619] The server uses the normalized data to train machine learning algorithms, which learn user preference patterns and improve prediction accuracy.

[0620] How it works: The server trains a collaborative filtering algorithm using the viewing history and rating data of multiple users with similar tastes.

[0621] Step 4:

[0622] Using the trained model, the server automatically recommends content based on the user's preferences, which can range from general popularity to niche content.

[0623] Specific behavior: The server analyzes that the user likes action movies and adds newly released action movies to the recommendation list.

[0624] Step 5:

[0625] The server collects word-of-mouth information about niche content and adds it to a recommendation list. This word-of-mouth information includes ratings and comments from other users.

[0626] Specific behavior: The server collects reviews of the movie "Drama Movie B" in which other users have rated it as "a moving story" and adds it to the recommendation list.

[0627] Step 6:

[0628] The server collects the user's emotion data using an emotion engine that recognizes the user's emotions, which includes the process of recognizing emotions from the user's facial expressions, voice, and text messages.

[0629] Specific operation: The server analyzes the user's facial expressions and tone of voice while watching the movie "Horror Movie C," and if it determines that the user is feeling scared, it stores that data.

[0630] Step 7:

[0631] The server optimizes the recommendation list based on the emotional data, adjusting the recommendations of relevant content according to the user's emotional state.

[0632] Specific behavior: The server recognizes that the user is prone to fear and generates a recommendation list that reduces the proportion of horror movies and increases the proportion of comedy movies.

[0633] Step 8:

[0634] The server transmits the generated recommendation list to the user terminal.

[0635] Specific operation: When a user opens the smartphone app, the server synchronizes with the local database and sends the latest recommendation list.

[0636] Step 9:

[0637] The terminal displays the received recommendation list to the user, who can then select and view content of interest from the list.

[0638] Specific operation: When a user opens a smartphone app, the device displays recommended movies and music on the home screen. The user can tap to watch.

[0639] Step 10:

[0640] Users can watch the recommended content and then add ratings and comments, which will be reflected in the next recommendations.

[0641] Specific behavior: A user watches the movie "Drama Movie B," rates it highly, and comments that "it had a touching ending." This information will be used to generate the next recommendation.

[0642] Step 11:

[0643] The server collects user feedback and stores it in a database, which is then used to generate recommendations next time.

[0644] What happens: The server stores the ratings and comments users leave for "Drama Movie B" in a database and uses them for the next algorithm training.

[0645] These processing steps enable highly accurate content recommendations based on the user's preferences and emotions.

[0646] Example 2

[0647] 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."

[0648] Conventional content recommendation systems rely on recommendation functions based on users' viewing history and rating data, but do not take into account the user's emotional state. This makes it difficult to provide optimal content to individual users, and there is a need to improve the user experience.

[0649] 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.

[0650] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing the recommendation list with word-of-mouth information, means for collecting emotional data to optimize the recommendation list, means for displaying the recommendation list on a user terminal, and means for collecting feedback, thereby enabling highly personalized recommendations that take into account not only the preferences but also the emotional state of each individual user.

[0651] "Means for collecting user data" refers to a function for collecting behavioral data such as user viewing history, ratings, and search history in real time or periodically.

[0652] "Means for normalizing collected data" refers to the ability to perform a process to convert data of different formats or scales into a consistent format.

[0653] The "means for training a machine learning algorithm" is a function for training a machine learning algorithm using normalized data to improve the accuracy of the model.

[0654] "Means for recommending content based on user preferences" is a function that uses a trained model to suggest content that best suits the user's preferences.

[0655] "Means for supplementing recommendation lists with word-of-mouth information" refers to a function that collects ratings and comments from other users and adds them to the recommendation list.

[0656] "Means of collecting emotional data to optimize recommendation lists" refers to a function that recognizes emotions from the user's facial expressions, voice, and text messages, and uses that data to adjust the recommendation list to be more optimal for the individual.

[0657] The "means for displaying the recommendation list on the user terminal" is a function for visually displaying the recommendation list sent from the server on the interface of the user terminal.

[0658] "Means for collecting feedback" is a function that collects ratings and comments from users, stores that information in a database, and reflects it in the next recommendation.

[0659] "Collaborative filtering" is a machine learning algorithm that recommends appropriate content to individual users based on the viewing history and rating data of multiple users with similar tastes.

[0660] This invention is a system that automatically recommends content based on user preferences and emotions. This system mainly consists of a server and a user terminal. The server is responsible for data collection, data normalization, machine learning algorithm training, emotion data collection, recommendation generation, and feedback collection. The user terminal provides an interface with the user and has the function of displaying a recommendation list.

[0661] User Data Collection

[0662] The server collects user behavioral data such as viewing history, ratings, and search history in real time or periodically. This allows the server to accurately grasp user preferences. For example, if a user watches "Action Movie A" and rates it 4 out of 5, the viewing history and rating are added to the database.

[0663] Data normalization

[0664] The server then processes the collected data to convert it into a consistent format, unifying data from different formats and scales, making it easier for machine learning algorithms to use. For example, data rated by a user on a 10-point scale is normalized to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0665] Training an AI model

[0666] The server trains machine learning algorithms based on the normalized data, which allows it to learn user preference patterns and improve prediction accuracy. For example, a collaborative filtering algorithm is used to learn the viewing history and rating data of multiple users with similar preferences.

[0667] Complementing content recommendations and word-of-mouth information

[0668] Using the trained model, the server recommends content based on the user's preferences. It also supplements the recommendation list with word-of-mouth information to provide other users' opinions and ratings. For example, it recommends "Action Movie A" and "Drama Movie B" to the user, along with other users' ratings and comments.

[0669] Combining Emotion Engines

[0670] The server uses an emotion engine to collect emotional data from the user's facial expressions, voice, and text messages, and optimizes the recommendation list. This process allows the server to provide content that best suits the user's current emotional state. For example, if the user expresses fear while watching "Horror Movie C," the server will select a less frightening movie for the next recommendation list.

[0671] Displaying the recommendation list

[0672] The user device uses a software interface that visually displays the recommendation list sent from the server, allowing the user to easily find new content. For example, when a user opens a smartphone app, the recommendation list is displayed on the home screen.

[0673] Gathering feedback

[0674] The server collects user ratings and comments in a database and reflects this information in the next recommendation. For example, if a user rates a movie they watched and leaves a comment saying "the story was good," that information will be used when generating the next recommendation.

[0675] In this way, the system of the present invention can realize highly accurate and personalized content recommendations based on the user's preferences and emotions, providing the user with a richer content experience.

[0676] Prompt Sentence Examples

[0677] "A user watched the movie 'Action Movie A' and rated it 4 out of 5. Please recommend new action movies based on their viewing history and ratings."

[0678] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0679] Step 1: Collect user data

[0680] The server collects user behavioral data, such as viewing history, ratings, and search history, in real time or periodically. Specifically, when a user watches a particular movie and enters a rating, the information is sent to the server. The input data is the movie title "Action Movie A" and the rating "4." The output is that these data are stored in a database.

[0681] Step 2: Normalize the data

[0682] The server then processes the collected raw user data to convert it into a consistent format. For example, it normalizes the data that the user has rated on a 10-point scale to a 0-1 scale. This unifies data from different formats and scales. The input data is a rating of 4, which is normalized to 0.8, and the normalized data is stored as the output.

[0683] Step 3: Training the AI ​​model

[0684] The server uses the normalized data to train a machine learning algorithm. Specifically, it uses a collaborative filtering algorithm to train the model on the viewing history and rating data of multiple users with similar preferences. The normalized rating data (0.8) and other users' data are used as input, and the output is the trained model.

[0685] Step 4: Recommend content and complement reviews

[0686] Using the trained model, the server recommends content based on the user's preferences. The server also supplements the recommendation list with reviews collected from other users. The input data is the trained model and review data, and the output is a recommendation list tailored to the user. Specifically, "Action Movie A" and "Drama Movie B" are included in the list, and reviews for each are displayed.

[0687] Step 5: Combining Emotion Engines

[0688] The server uses an emotion engine to collect user emotional data and optimize the recommendation list. It recognizes emotions from the user's facial expressions, voice, and text messages and makes adjustments based on that data. The input data is the user's emotional data while watching, and the output is a recommendation list that takes emotions into consideration. Specifically, if the server recognizes that the user is feeling scared, the next movie recommendation will be a less scary one.

[0689] Step 6: Displaying the recommendation list

[0690] The device uses a software interface that visually presents the recommendation list sent from the server. The input is the recommendation list data from the server, and the output is the list displayed on the device screen. When a user opens the smartphone app, a list of related recommended movies and TV shows is displayed on the home screen.

[0691] Step 7: Gather feedback

[0692] The server collects ratings and comments from users and reflects this information in the next recommendation. The input data are user ratings and comments, which are saved in a database and used to generate the next recommendation. The output is an updated database. For example, if a user rates a story as "good," that information is reflected in the server.

[0693] (Application example 2)

[0694] 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."

[0695] Conventional content recommendation systems make recommendations based on user preferences and behavioral data, but do not take into account the user's real-time emotions when recommending content. As a result, content that does not match the user's momentary emotions or state may be recommended, resulting in a poor user experience. The present invention aims to provide a system that combines the user's real-time emotional information to realize highly accurate and personalized content recommendations.

[0696] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0697] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing a recommendation list with word-of-mouth information, means for displaying the recommendation list on a user terminal, means for acquiring emotion data using the user's facial expressions and voice, and means for adjusting the recommendation list in real time based on the emotion data. This makes it possible to recommend more appropriate content based on the user's real-time emotions.

[0698] "User data" refers to information about a user's behavior and preferences, such as their viewing history, ratings, and search history.

[0699] "Means for collecting" refers to a method or device by which the server collects behavioral data such as user viewing history and ratings in real time or periodically.

[0700] "Normalization means" refers to a method or device that converts collected data into a consistent format and unifies data of different formats or scales.

[0701] "Means for training" refers to a method or apparatus that uses normalized data to train a machine learning algorithm.

[0702] "Means for recommending" refers to a method or apparatus that uses a trained model to suggest content based on a user's preferences.

[0703] The "means for supplementing word-of-mouth information" refers to a method or device for collecting opinions and ratings from other users and adding them to a recommendation list.

[0704] The "means for displaying" refers to a software interface that visually presents the recommendation list on the user terminal.

[0705] "Means for acquiring emotion data using facial expressions and voice" refers to a method or device for capturing a user's facial expressions and voice in real time and recognizing their emotions.

[0706] "Means for adjusting a recommendation list in real time" refers to a method or apparatus for optimizing a recommendation list in real time based on acquired emotion data.

[0707] "Machine learning algorithm" refers to an algorithm that learns user preference patterns based on collected data and improves prediction accuracy.

[0708] "Collaborative filtering" refers to a technique for improving recommendation accuracy by using data from multiple users with similar preferences.

[0709] The present invention is a system for automatically recommending content based on a user's preferences and emotions, and an embodiment thereof will be described in detail below.

[0710] System configuration

[0711] The system includes means for collecting user data, normalizing the data, training machine learning algorithms, recommending content, complementing reviews, displaying recommendation lists, acquiring sentiment data, and adjusting recommendations in real time.

[0712] Data collection

[0713] The server collects user behavioral data, such as viewing history, ratings, and search history, in real time or periodically. To do so, it obtains the user's viewing history and ratings from a database. It also collects facial expressions and voice data using the smartphone's camera and microphone.

[0714] Data normalization

[0715] The server converts the collected data into a consistent format, specifically normalizing the rating data to a scale of 0 to 1. This unifies the variations in data format and scale, making it easier for algorithms to use.

[0716] Training machine learning algorithms

[0717] The server uses the normalized data to train a machine learning algorithm, using a collaborative filtering algorithm to build a model using data from multiple users with similar preferences.

[0718] Content recommendations

[0719] The trained model is used to predict user preference patterns and recommend the most suitable content, thereby providing personalized content to users.

[0720] Complementing reviews

[0721] The server collects word-of-mouth information from other users and adds it to the recommendation list, making it easier for users to refer to the opinions and ratings of other users.

[0722] Displaying the recommendation list

[0723] The recommendation list sent from the server is visually displayed on the user's device, allowing the user to select and view content of interest from the list.

[0724] Acquiring emotion data

[0725] The server uses the smartphone's camera and microphone to obtain emotion data from the user's facial expressions and voice, using a neural network model for emotion analysis.

[0726] Real-time recommendation adjustment

[0727] The server adjusts the recommendation list in real time based on the acquired emotional data. For example, if a user is watching a scary movie, the server will recognize that emotional data in real time and recommend less scary content next.

[0728] Specific examples

[0729] If a user feels scared while watching "Horror Movie C," the server will recognize that emotion in real time and recommend a less scary movie for the next viewing, taking into account word-of-mouth reviews from other users to suggest more appropriate content.

[0730] Prompt Sentence Examples

[0731] "If it is detected in the user's viewing history that they felt scared while watching 'Horror Movie C,' recommend movies that are less scary based on their viewing history and rating data. Also, take into account word-of-mouth information from other users."

[0732] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0733] Step 1:

[0734] User Data Collection

[0735] Input: User's viewing history, ratings, search history, facial expression and voice data from the smartphone camera and microphone

[0736] The server collects user behavior data from a database in real time or periodically, and also collects facial expression and voice data using the smartphone's camera and microphone.

[0737] Output: Collected user viewing history, rating data, search history, facial expression data, and voice data

[0738] Step 2:

[0739] Data normalization

[0740] Input: Collected viewing history, rating data, search history

[0741] The server normalizes the collected rating data to a scale of 0 to 1. Specifically, if the rating is out of 10, it converts it to a scale of 0 to 1.

[0742] Output: Normalized rating data

[0743] Step 3:

[0744] Training a machine learning model

[0745] Input: Normalized rating data, viewing history, search history

[0746] The server uses the normalized data to train a collaborative filtering algorithm, building a model using data from multiple users with similar preferences.

[0747] Output: A trained machine learning model

[0748] Step 4:

[0749] Content recommendations

[0750] Input: Trained machine learning model, current user preference data

[0751] The server uses the trained model to recommend optimal content based on the current user preferences.

[0752] Output: Recommendation list

[0753] Step 5:

[0754] Complementing reviews

[0755] Input: Recommendation list, word-of-mouth data from other users

[0756] The server adds word-of-mouth information from other users to the recommendation list, thereby providing other users' opinions and ratings about the recommended content.

[0757] Output: Recommendation list with reviews added

[0758] Step 6:

[0759] Displaying the recommendation list

[0760] Input: Recommendation list with reviews

[0761] The server visually provides a recommendation list to the user's device, allowing the user to select and view content of interest from the list.

[0762] Output: Recommendation list displayed on the user's device

[0763] Step 7:

[0764] Acquiring emotion data

[0765] Input: User's facial expression data, voice data

[0766] The server uses the smartphone's camera and microphone to analyze data obtained from the user's facial expressions and voice using an emotion recognition algorithm, and uses a neural network to estimate the user's emotions.

[0767] Output: Real-time emotion data

[0768] Step 8:

[0769] Real-time recommendation adjustment

[0770] Input: Real-time emotion data, recommendation list

[0771] The server adjusts the recommendation list based on real-time emotion data. For example, if a user feels scared, the server will reflect that emotion and recommend a less scary movie as the next content to watch.

[0772] Output: A refined list of recommendations

[0773] 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.

[0774] 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.

[0775] 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.

[0776] [Third embodiment]

[0777] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0778] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0779] 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).

[0780] 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.

[0781] 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.

[0782] 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).

[0783] 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. 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.

[0784] 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.

[0785] 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.

[0786] 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.

[0787] 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.

[0788] 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."

[0789] The present invention is a system for automatically recommending content based on a user's preferences, and an embodiment thereof will be described in detail below.

[0790] User Data Collection

[0791] The means of collection is for the server to collect data such as users' viewing history, ratings, search history, etc. in real time or periodically. This is done based on the content viewed by the user and the ratings given.

[0792] Example: A server automatically records viewing history data and receives a rating score each time a user watches a movie.

[0793] Data normalization

[0794] Normalization is the process of converting collected data into a consistent format, making data of different formats and scales uniform and easier for algorithms to use.

[0795] Example: The server converts data rated by users on different scales (e.g., 10-point or 5-point) into a scale from 0 to 1.

[0796] Training an AI model

[0797] Next, a training means is used to train the machine learning algorithm based on the collected data, thereby learning user preference patterns and improving prediction accuracy.

[0798] Example: A server trains a collaborative filtering algorithm using data from a group of users with similar tastes to build a content recommendation model tailored to each user's preferences.

[0799] Complementing content recommendations and word-of-mouth information

[0800] Using the trained model, the system proposes content based on user preferences through a recommendation tool. It also collects user reviews and adds them to the recommendation list, allowing users to easily refer to other users' opinions and ratings.

[0801] Example: The server recommends two movies, "Action Movie A" and "Drama Movie B," based on the user's preferences, and adds evaluation comments from other users.

[0802] Displaying the recommendation list

[0803] The display method is a software interface that visually displays the recommendation list on the user's device, allowing the user to easily find new content.

[0804] Example: When a user opens a smartphone app, a recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[0805] Gathering feedback

[0806] Finally, we use a means to collect feedback, and store user ratings and comments in a database, which can be reflected in future recommendations to improve accuracy.

[0807] Example: If a user rates a movie they watched and leaves a comment saying "the story was good," the server collects that information and uses it for the next recommendation.

[0808] This allows the system of the present invention to provide a richer content experience to the user.

[0809] The processing flow will be explained below.

[0810] Step 1:

[0811] The server collects real-time user behavioral data such as viewing history, ratings, and search history. Every time a user watches a new movie or adds a rating, the information is stored in a database.

[0812] Specific operation: When a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[0813] Step 2:

[0814] The server normalizes the collected data, which involves converting data measured at different scales into a unified format.

[0815] Specific behavior: The server normalizes the user's rating data from a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0816] Step 3:

[0817] The server uses the normalized data to train machine learning algorithms, which learn user preference patterns and improve prediction accuracy.

[0818] How it works: The server trains a collaborative filtering algorithm using the viewing history and rating data of multiple users with similar tastes.

[0819] Step 4:

[0820] Using the trained model, the server automatically recommends content based on the user's preferences, which can range from general popularity to niche content.

[0821] Specific behavior: The server analyzes that the user likes action movies and adds newly released action movies to the recommendation list.

[0822] Step 5:

[0823] The server collects word-of-mouth information about niche content and adds it to a recommendation list. This word-of-mouth information includes ratings and comments from other users.

[0824] Specific behavior: The server collects reviews of the movie "Drama Movie B" in which other users have rated it as "a moving story" and adds it to the recommendation list.

[0825] Step 6:

[0826] The server transmits the generated recommendation list to the user terminal.

[0827] Specific operation: When a user opens the smartphone app, the server synchronizes with the local database and sends the latest recommendation list.

[0828] Step 7:

[0829] The device displays the received recommendation list to the user, who can then select and view content of interest from the list.

[0830] Specific operation: When a user opens a smartphone app, the device displays recommended movies and music on the home screen. The user can tap to watch.

[0831] Step 8:

[0832] Users can watch the recommended content and then add ratings and comments, which will be reflected in the next recommendations.

[0833] Specific behavior: A user watches the movie "Drama Movie B," rates it highly, and comments that "it had a touching ending." This information will be used to generate the next recommendation.

[0834] Step 9:

[0835] The server collects user feedback and stores it in a database, which is then used to generate recommendations next time.

[0836] What happens: The server stores the ratings and comments users leave for "Drama Movie B" in a database and uses them for the next algorithm training.

[0837] These processing steps enable highly accurate content recommendations based on the user's preferences.

[0838] Example 1

[0839] 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."

[0840] Current content recommendation systems often lack sufficient accuracy in recommending content based on user preferences. They also often fail to provide effective recommendations to new users. Furthermore, they lack a mechanism for utilizing word-of-mouth information or efficiently collecting user feedback, which means they are unable to respond to real-time changes in user preferences. This can result in a poor user experience and a deterioration in overall system performance.

[0841] 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.

[0842] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing the recommendation list with word-of-mouth information, means for displaying the recommendation list on the user's terminal, means for collecting the user's viewing history from the terminal and transmitting it to the server, means for converting the collected data into a consistent format, and means for collecting feedback data based on user ratings and storing it in a database. This allows for accurate understanding of user preferences and enables effective content recommendations even for new users. Furthermore, utilizing word-of-mouth information increases user satisfaction, and collecting feedback continuously improves the accuracy of the system and the user experience.

[0843] "User data" refers to information that indicates a user's behavior and preferences, such as the user's viewing history, ratings, and search history.

[0844] "Collection Methods" refers to the technologies and processes used to obtain and transmit User Data to a server.

[0845] "Normalization methods" refer to techniques and processes that convert collected data into a consistent format and make it easier for algorithms to use.

[0846] A "machine learning algorithm" refers to a computer program that learns patterns and rules from data to make predictions and classifications.

[0847] "Training" refers to the process of training a machine learning algorithm using collected data.

[0848] "Recommendation method" refers to the technology or process that uses a trained model to recommend content based on user preferences.

[0849] "Word-of-mouth information" refers to evaluation comments and reviews of content left by other users.

[0850] A "recommendation list" refers to a list of content recommended to a user.

[0851] "Means for displaying" refers to the technology or process for visually presenting the recommendation list to the user's terminal.

[0852] "User terminal" refers to the device a user uses to interact with the system, such as a smartphone, PC, or tablet.

[0853] "Viewing history" refers to information about content that a user has viewed in the past.

[0854] "Consistent format" refers to a data format that unifies data of different scales and formats and makes it easy for algorithms to use.

[0855] "Feedback data" refers to the rating scores and comments provided by users after viewing content.

[0856] "Database" refers to a system for systematically storing and managing collected user data and feedback data.

[0857] The present invention is a system that automatically recommends content based on a user's preferences, and details thereof will be described below.

[0858] System Overview

[0859] The system mainly consists of the following elements:

[0860] How user data is collected

[0861] A means of normalizing data

[0862] A means of training machine learning algorithms

[0863] Recommendation methods

[0864] A means to complement word-of-mouth information

[0865] How to display a recommendation list

[0866] A means of gathering feedback

[0867] User Data Collection

[0868] The server collects users' viewing history, ratings, and search history in real time or periodically. The data includes the ID, start time, end time, and rating of the content viewed by the user. This data is stored in the server's database and used for analysis.

[0869] Example: When a user watches the movie "Action Movie A," the viewing history is stored in a database on the server.

[0870] Data normalization

[0871] The server converts the collected data into a consistent format, converting ratings from different scales (e.g., 10-point or 5-point scales) into a range between 0 and 1, and standardizing data in other formats, making the data consistent and suitable for training algorithms.

[0872] Example: Ratings of "4 out of 5" and "8 out of 10" are converted to the format "0.8".

[0873] Training an AI model

[0874] The server uses the normalized data to train machine learning algorithms, using techniques such as collaborative filtering and deep learning to learn user preference patterns, resulting in the creation of highly accurate recommendation models.

[0875] Example: The server uses past viewing and rating data to learn each user's preference patterns.

[0876] Complementing content recommendations and word-of-mouth information

[0877] The server uses the trained model to recommend content based on the user's preferences, and adds word-of-mouth information collected from other users to the recommendation list, making it easier for users to refer to other users' ratings.

[0878] Example: Recommend the movies "Action Movie A" and "Drama Movie B" and add ratings and comments from other users.

[0879] Displaying the recommendation list

[0880] The user device (smartphone, PC, tablet, etc.) visually displays the recommendation list sent from the server. When the user opens the application, they can select and watch content of their interest from the recommendation list displayed on the home screen.

[0881] Example: "Action Movie A" and "Drama Movie B" are displayed on the home screen.

[0882] Gathering feedback

[0883] The server stores user feedback (ratings and comments) in a database, which is used to improve the accuracy of recommendations next time.

[0884] Example: If a user watches the movie "Drama Movie B" and then comments that "the story was good," that information is saved on the server and used for the next recommendation.

[0885] Example prompts to input to the generative AI model

[0886] "Please enter the title and rating of the movie you watched:

[0887] Movie Title: "Action Movie A"

[0888] Review comment: "The action scenes were powerful."

[0889] Use this information to generate an up-to-date list of recommendations for your users."

[0890] This system enables highly accurate and personalized content recommendations to users, improving the user experience.

[0891] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0892] Step 1: Collect user data

[0893] Input: Information about the content viewed by the user (ID, viewing start time, end time, rating score)

[0894] How it works: Every time a user watches content, their device automatically sends viewing data to the server, which then records the data in a database. If the user enters a rating after viewing, the rating data is also sent to the server.

[0895] Output: Viewing history and rating data stored in a database

[0896] Step 2: Normalize the data

[0897] Input: Raw data in the database (viewing history, rating data, search history)

[0898] Specific operation: The server performs normalization processing on the collected data. For example, it standardizes all data rated on different scales (10-point or 5-point) to a scale from 0 to 1.

[0899] Output: Normalized, consistent data

[0900] Step 3: Training the AI ​​model

[0901] Input: Normalized viewing history data and rating data

[0902] Specific operation: The server uses the normalized data to train machine learning algorithms (e.g., collaborative filtering and deep learning), extracting features from the data and analyzing user preference patterns to build a model.

[0903] Output: A trained, highly accurate recommendation model

[0904] Step 4: Recommend content and complement reviews

[0905] Input: Trained AI model and customer reviews

[0906] Specific operation: The server uses the trained model to generate a content list optimized for each user. At the same time, it retrieves other users' evaluation comments from a word-of-mouth information database and adds them to the recommendation list.

[0907] Output: Recommendation list supplemented with reviews

[0908] Step 5: View the recommendation list

[0909] Input: Recommendation list sent from the server

[0910] Specific operation: When a user opens the application on a device (such as a smartphone or PC), a recommendation list sent from the server is displayed. The user can select and watch content that interests them from this list.

[0911] Output: Recommendation list displayed on the user's device

[0912] Step 6: Gather feedback

[0913] Input: Rating score and comments entered by users after viewing the content

[0914] How it works: After watching a video, users can leave a rating or comment, and that information is sent from the device to the server, which then stores this feedback in a database and uses it to train the recommendation model next time.

[0915] Output: User feedback data stored in a database

[0916] Through the above processing steps, the system can provide highly accurate content based on the user's preferences.

[0917] (Application example 1)

[0918] 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."

[0919] In today's world, it is extremely difficult for users to find content that suits their preferences from the vast amount of content available. Furthermore, conventional content recommendation systems are inadequately responsive to the specific devices and platforms of individual users. Furthermore, few systems offer a function that allows users to easily refer to word-of-mouth information, leaving users with a lack of reference information when making their selections.

[0920] 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.

[0921] In this invention, the server includes means for collecting user data, means for normalizing the data, means for training with a machine learning algorithm, means for recommending content based on user preferences using a smartphone, and means for supplementing word-of-mouth information. This allows users to easily find content that matches their preferences on their smartphone, and by referring to word-of-mouth information from other users, they can select content that will provide them with greater satisfaction.

[0922] "User data" refers to information such as a user's viewing history, ratings, and search history, and is data that indicates the user's preferences and behavior.

[0923] "Normalization" is the process of converting collected data into a consistent format and unifying data of different scales and formats.

[0924] A "machine learning algorithm" is a technology that analyzes data, learns user behavior patterns and preferences, and makes predictions and classifications.

[0925] "Recommendation" refers to selecting suitable content based on a user's preferences and presenting it to the user.

[0926] "Word-of-mouth information" refers to information such as ratings and comments provided by other users, and is opinions and impressions that can be used as a reference for making a selection.

[0927] A "user terminal" is an electronic device used by a user to display information and perform operations.

[0928] "Feedback" refers to reaction data such as ratings and comments provided by users after use.

[0929] "Collaborative filtering" is a method for recommending optimal content to individual users by comparing their preferences with data from other users who have similar preferences.

[0930] A "smartphone" is a portable information terminal with multiple functions, including communication, information processing, and application execution.

[0931] A "personalized movie recommendation app" is an application that recommends movies and TV dramas based on the preferences of each individual user.

[0932] The present invention is a personalized recommendation system that enables users to efficiently find content such as movies and dramas on their smartphones. The system of the present invention is configured as follows.

[0933] User Data Collection

[0934] The server collects data such as users' viewing history, ratings, and search history in real time or periodically. This provides basic data for understanding users' preferences. Specifically, each time a user watches a movie or drama, the server records the user's viewing history and rating score. Search history is also collected to understand the genres and titles in which the user is interested.

[0935] Data normalization

[0936] The server normalizes the collected data. User ratings are often rated on different scales (e.g., 10-point scale or 5-point scale). This data is converted to a unified scale (ranging from 0 to 1) to make it easier for machine learning algorithms to use.

[0937] Training an AI model

[0938] Using the collected and normalized data, the server trains a machine learning algorithm. This uses a collaborative filtering algorithm to build a content recommendation model for each user based on data from users with similar tastes. The trained model learns the user's preference patterns and can make more accurate recommendations.

[0939] Complementing content recommendations and word-of-mouth information

[0940] Using the trained model, the server recommends movies and TV shows based on the user's preferences to the smartphone. It also collects word-of-mouth information from other users and adds it to the recommendation list, allowing users to refer to other users' opinions and ratings.

[0941] Displaying the recommendation list

[0942] When a user opens the smartphone app, the recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[0943] Gathering feedback

[0944] When users leave ratings and comments on movies and TV shows they have watched, the server collects that feedback and stores it in a database. This information is then reflected in the next recommendation, further improving the system's prediction accuracy.

[0945] Through the above process, the system can provide the user with a personalized content experience. As a specific example, the following prompt sentence can be used:

[0946] Example prompt sentence:

[0947] Based on user ID 1's recent viewing history, what movies would you recommend to him?

[0948] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0949] Step 1: Collect user data

[0950] The server collects data such as the user's viewing history, rating scores, and search history. This data is collected in real time or periodically. For example, every time a user watches a movie, the viewing history and rating score are sent to the server and recorded in a database. The input is the user's viewing history, rating scores, and search history, and the output is the consistently recorded user data.

[0951] Step 2: Normalize the data

[0952] The server normalizes the collected user data into a consistent format. Specifically, it converts data rated on different scales into a scale from 0 to 1. This process unifies the data and makes it easier to use in machine learning algorithms. The input is the rating data on different scales, and the output is data normalized from 0 to 1.

[0953] Step 3: Training the AI ​​model

[0954] The server uses the normalized data to train a machine learning algorithm, which uses a collaborative filtering algorithm to build a content recommendation model based on data from a group of users with similar preferences. The input is the normalized user data, and the output is the trained recommendation model.

[0955] Step 4: Recommend content and complement reviews

[0956] The server uses the trained model to recommend content based on the user's preferences. In addition, word-of-mouth information collected from other users is added to the recommendation list, allowing users to refer to other users' ratings and comments. The input is the trained recommendation model and word-of-mouth information, and the output is a recommendation list with the word-of-mouth information added.

[0957] Step 5: View the recommendation list

[0958] When a user opens the smartphone app, the server sends a recommendation list to the user's device and displays it on the home screen. The user can select and watch content of interest from this list. The input is a recommendation list with added word-of-mouth information, and the output is a recommendation list displayed on the user's smartphone.

[0959] Step 6: Gather feedback

[0960] When users leave ratings or comments on movies or TV shows they have watched, the server collects that feedback and stores it in a database. The feedback data is reflected in the next recommendation. The input is the user's ratings and comments, and the output is the feedback data stored in the database.

[0961] The above is a specific processing flow in the system of the present invention.

[0962] 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.

[0963] The present invention is a system for automatically recommending content based on a user's preferences and emotions, and an embodiment thereof will be described in detail below.

[0964] User Data Collection

[0965] The means of collection is for the server to collect behavioral data such as user viewing history, ratings, and search history in real time or periodically, based on the content viewed by the user and the ratings given.

[0966] Example: If a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[0967] Data normalization

[0968] Normalization is the process of converting collected data into a consistent format, making data of different formats and scales uniform and easier for algorithms to use.

[0969] Example: The server normalizes the data that the user has rated on a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0970] Training an AI model

[0971] Next, a training means is used to train the machine learning algorithm based on the collected data, thereby learning user preference patterns and improving prediction accuracy.

[0972] Example: A server trains a collaborative filtering algorithm using viewing history and rating data from multiple users with similar tastes.

[0973] Complementing content recommendations and word-of-mouth information

[0974] Using the trained model, the system proposes content based on user preferences through a recommendation tool. It also collects user reviews and adds them to the recommendation list, allowing users to easily refer to other users' opinions and ratings.

[0975] Example: The server recommends two movies, "Action Movie A" and "Drama Movie B," based on the user's preferences, and adds evaluation comments from other users.

[0976] Combining Emotion Engines

[0977] Furthermore, it combines an emotion engine to collect user emotional data, which recognizes emotions from users' facial expressions, voice, and text messages, and optimizes the recommendation list based on that data.

[0978] Example: The server monitors the user's facial expressions and tone of voice while watching the movie "Horror Movie C." If the server determines that the user is feeling scared, it will select a less frightening movie for the next recommendation.

[0979] Displaying the recommendation list

[0980] The display method is a software interface that visually displays the recommendation list on the user's device, allowing the user to easily find new content.

[0981] Example: When a user opens a smartphone app, a recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[0982] Gathering feedback

[0983] Finally, we use a feedback collection method to store user ratings and comments in a database, which can be used to improve the accuracy of future recommendations.

[0984] Example: If a user rates a movie they watched and leaves a comment saying "the story was good," the server collects that information and uses it when generating the next recommendation.

[0985] In this way, the system of the present invention can realize highly accurate and personalized content recommendations based on the user's preferences and emotions, and provide the user with a richer content experience.

[0986] The processing flow will be explained below.

[0987] Step 1:

[0988] The server collects real-time user behavioral data such as viewing history, ratings, and search history. Every time a user watches a new movie or adds a rating, the information is stored in a database.

[0989] Specific operation: When a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[0990] Step 2:

[0991] The server normalizes the collected data, which involves converting data measured at different scales into a unified format.

[0992] Specific behavior: The server normalizes the user's rating data from a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[0993] Step 3:

[0994] The server uses the normalized data to train machine learning algorithms, which learn user preference patterns and improve prediction accuracy.

[0995] How it works: The server trains a collaborative filtering algorithm using the viewing history and rating data of multiple users with similar tastes.

[0996] Step 4:

[0997] Using the trained model, the server automatically recommends content based on the user's preferences, which can range from general popularity to niche content.

[0998] Specific behavior: The server analyzes that the user likes action movies and adds newly released action movies to the recommendation list.

[0999] Step 5:

[1000] The server collects word-of-mouth information about niche content and adds it to a recommendation list. This word-of-mouth information includes ratings and comments from other users.

[1001] Specific behavior: The server collects reviews of the movie "Drama Movie B" in which other users have rated it as "a moving story" and adds it to the recommendation list.

[1002] Step 6:

[1003] The server collects the user's emotion data using an emotion engine that recognizes the user's emotions, which includes the process of recognizing emotions from the user's facial expressions, voice, and text messages.

[1004] Specific operation: The server analyzes the user's facial expressions and tone of voice while watching the movie "Horror Movie C," and if it determines that the user is feeling scared, it stores that data.

[1005] Step 7:

[1006] The server optimizes the recommendation list based on the emotional data, adjusting the recommendations of relevant content according to the user's emotional state.

[1007] Specific behavior: The server recognizes that the user is prone to fear and generates a recommendation list that reduces the proportion of horror movies and increases the proportion of comedy movies.

[1008] Step 8:

[1009] The server transmits the generated recommendation list to the user terminal.

[1010] Specific operation: When a user opens the smartphone app, the server synchronizes with the local database and sends the latest recommendation list.

[1011] Step 9:

[1012] The terminal displays the received recommendation list to the user, who can then select and view content of interest from the list.

[1013] Specific operation: When a user opens a smartphone app, the device displays recommended movies and music on the home screen. The user can tap to watch.

[1014] Step 10:

[1015] Users can watch the recommended content and then add ratings and comments, which will be reflected in the next recommendations.

[1016] Specific behavior: A user watches the movie "Drama Movie B," rates it highly, and comments that "it had a touching ending." This information will be used to generate the next recommendation.

[1017] Step 11:

[1018] The server collects user feedback and stores it in a database, which is then used to generate recommendations next time.

[1019] What happens: The server stores the ratings and comments users leave for "Drama Movie B" in a database and uses them for the next algorithm training.

[1020] These processing steps enable highly accurate content recommendations based on the user's preferences and emotions.

[1021] Example 2

[1022] 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."

[1023] Conventional content recommendation systems rely on recommendation functions based on users' viewing history and rating data, but do not take into account the user's emotional state. This makes it difficult to provide optimal content to individual users, and there is a need to improve the user experience.

[1024] 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.

[1025] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing the recommendation list with word-of-mouth information, means for collecting emotional data to optimize the recommendation list, means for displaying the recommendation list on a user terminal, and means for collecting feedback, thereby enabling highly personalized recommendations that take into account not only the preferences but also the emotional state of each individual user.

[1026] "Means for collecting user data" refers to a function for collecting behavioral data such as user viewing history, ratings, and search history in real time or periodically.

[1027] "Means for normalizing collected data" refers to the ability to perform a process to convert data of different formats or scales into a consistent format.

[1028] The "means for training a machine learning algorithm" is a function for training a machine learning algorithm using normalized data to improve the accuracy of the model.

[1029] "Means for recommending content based on user preferences" is a function that uses a trained model to suggest content that best suits the user's preferences.

[1030] "Means for supplementing recommendation lists with word-of-mouth information" refers to a function that collects ratings and comments from other users and adds them to the recommendation list.

[1031] "Means of collecting emotional data to optimize recommendation lists" refers to a function that recognizes emotions from the user's facial expressions, voice, and text messages, and uses that data to adjust the recommendation list to be more optimal for the individual.

[1032] The "means for displaying the recommendation list on the user terminal" is a function for visually displaying the recommendation list sent from the server on the interface of the user terminal.

[1033] "Means for collecting feedback" is a function that collects ratings and comments from users, stores that information in a database, and reflects it in the next recommendation.

[1034] "Collaborative filtering" is a machine learning algorithm that recommends appropriate content to individual users based on the viewing history and rating data of multiple users with similar tastes.

[1035] This invention is a system that automatically recommends content based on user preferences and emotions. This system mainly consists of a server and a user terminal. The server is responsible for data collection, data normalization, machine learning algorithm training, emotion data collection, recommendation generation, and feedback collection. The user terminal provides an interface with the user and has the function of displaying a recommendation list.

[1036] User Data Collection

[1037] The server collects user behavioral data such as viewing history, ratings, and search history in real time or periodically. This allows the server to accurately grasp user preferences. For example, if a user watches "Action Movie A" and rates it 4 out of 5, the viewing history and rating are added to the database.

[1038] Data normalization

[1039] The server then processes the collected data to convert it into a consistent format, unifying data from different formats and scales, making it easier for machine learning algorithms to use. For example, data rated by a user on a 10-point scale is normalized to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[1040] Training an AI model

[1041] The server trains machine learning algorithms based on the normalized data, which allows it to learn user preference patterns and improve prediction accuracy. For example, a collaborative filtering algorithm is used to learn the viewing history and rating data of multiple users with similar preferences.

[1042] Complementing content recommendations and word-of-mouth information

[1043] Using the trained model, the server recommends content based on the user's preferences. It also supplements the recommendation list with word-of-mouth information to provide other users' opinions and ratings. For example, it recommends "Action Movie A" and "Drama Movie B" to the user, along with other users' ratings and comments.

[1044] Combining Emotion Engines

[1045] The server uses an emotion engine to collect emotional data from the user's facial expressions, voice, and text messages, and optimizes the recommendation list. This process allows the server to provide content that best suits the user's current emotional state. For example, if the user expresses fear while watching "Horror Movie C," the server will select a less frightening movie for the next recommendation list.

[1046] Displaying the recommendation list

[1047] The user device uses a software interface that visually displays the recommendation list sent from the server, allowing the user to easily find new content. For example, when a user opens a smartphone app, the recommendation list is displayed on the home screen.

[1048] Gathering feedback

[1049] The server collects user ratings and comments in a database and reflects this information in the next recommendation. For example, if a user rates a movie they watched and leaves a comment saying "the story was good," that information will be used when generating the next recommendation.

[1050] In this way, the system of the present invention can realize highly accurate and personalized content recommendations based on the user's preferences and emotions, providing the user with a richer content experience.

[1051] Prompt Sentence Examples

[1052] "A user watched the movie 'Action Movie A' and rated it 4 out of 5. Please recommend new action movies based on their viewing history and ratings."

[1053] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1054] Step 1: Collect user data

[1055] The server collects user behavioral data, such as viewing history, ratings, and search history, in real time or periodically. Specifically, when a user watches a particular movie and enters a rating, the information is sent to the server. The input data is the movie title "Action Movie A" and the rating "4." The output is that these data are stored in a database.

[1056] Step 2: Normalize the data

[1057] The server then processes the collected raw user data to convert it into a consistent format. For example, it normalizes the data that the user has rated on a 10-point scale to a 0-1 scale. This unifies data from different formats and scales. The input data is a rating of 4, which is normalized to 0.8, and the normalized data is stored as the output.

[1058] Step 3: Training the AI ​​model

[1059] The server uses the normalized data to train a machine learning algorithm. Specifically, it uses a collaborative filtering algorithm to train the model on the viewing history and rating data of multiple users with similar preferences. The normalized rating data (0.8) and other users' data are used as input, and the output is the trained model.

[1060] Step 4: Recommend content and complement reviews

[1061] Using the trained model, the server recommends content based on the user's preferences. The server also supplements the recommendation list with reviews collected from other users. The input data is the trained model and review data, and the output is a recommendation list tailored to the user. Specifically, "Action Movie A" and "Drama Movie B" are included in the list, and reviews for each are displayed.

[1062] Step 5: Combining Emotion Engines

[1063] The server uses an emotion engine to collect user emotional data and optimize the recommendation list. It recognizes emotions from the user's facial expressions, voice, and text messages and makes adjustments based on that data. The input data is the user's emotional data while watching, and the output is a recommendation list that takes emotions into consideration. Specifically, if the server recognizes that the user is feeling scared, the next movie recommendation will be a less scary one.

[1064] Step 6: Displaying the recommendation list

[1065] The device uses a software interface that visually presents the recommendation list sent from the server. The input is the recommendation list data from the server, and the output is the list displayed on the device screen. When a user opens the smartphone app, a list of related recommended movies and TV shows is displayed on the home screen.

[1066] Step 7: Gather feedback

[1067] The server collects ratings and comments from users and reflects this information in the next recommendation. The input data are user ratings and comments, which are saved in a database and used to generate the next recommendation. The output is an updated database. For example, if a user rates a story as "good," that information is reflected in the server.

[1068] (Application example 2)

[1069] 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."

[1070] Conventional content recommendation systems make recommendations based on user preferences and behavioral data, but do not take into account the user's real-time emotions when recommending content. As a result, content that does not match the user's momentary emotions or state may be recommended, resulting in a poor user experience. The present invention aims to provide a system that combines the user's real-time emotional information to realize highly accurate and personalized content recommendations.

[1071] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1072] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing a recommendation list with word-of-mouth information, means for displaying the recommendation list on a user terminal, means for acquiring emotion data using the user's facial expressions and voice, and means for adjusting the recommendation list in real time based on the emotion data. This makes it possible to recommend more appropriate content based on the user's real-time emotions.

[1073] "User data" refers to information about a user's behavior and preferences, such as their viewing history, ratings, and search history.

[1074] "Means for collecting" refers to a method or device by which the server collects behavioral data such as user viewing history and ratings in real time or periodically.

[1075] "Normalization means" refers to a method or device that converts collected data into a consistent format and unifies data of different formats or scales.

[1076] "Means for training" refers to a method or apparatus that uses normalized data to train a machine learning algorithm.

[1077] "Means for recommending" refers to a method or apparatus that uses a trained model to suggest content based on a user's preferences.

[1078] The "means for supplementing word-of-mouth information" refers to a method or device for collecting opinions and ratings from other users and adding them to a recommendation list.

[1079] The "means for displaying" refers to a software interface that visually presents the recommendation list on the user terminal.

[1080] "Means for acquiring emotion data using facial expressions and voice" refers to a method or device for capturing a user's facial expressions and voice in real time and recognizing their emotions.

[1081] "Means for adjusting a recommendation list in real time" refers to a method or apparatus for optimizing a recommendation list in real time based on acquired emotion data.

[1082] "Machine learning algorithm" refers to an algorithm that learns user preference patterns based on collected data and improves prediction accuracy.

[1083] "Collaborative filtering" refers to a technique for improving recommendation accuracy by using data from multiple users with similar preferences.

[1084] The present invention is a system for automatically recommending content based on a user's preferences and emotions, and an embodiment thereof will be described in detail below.

[1085] System configuration

[1086] The system includes means for collecting user data, normalizing the data, training machine learning algorithms, recommending content, complementing reviews, displaying recommendation lists, acquiring sentiment data, and adjusting recommendations in real time.

[1087] Data collection

[1088] The server collects user behavioral data, such as viewing history, ratings, and search history, in real time or periodically. To do so, it obtains the user's viewing history and ratings from a database. It also collects facial expressions and voice data using the smartphone's camera and microphone.

[1089] Data normalization

[1090] The server converts the collected data into a consistent format, specifically normalizing the rating data to a scale of 0 to 1. This unifies the variations in data format and scale, making it easier for algorithms to use.

[1091] Training machine learning algorithms

[1092] The server uses the normalized data to train a machine learning algorithm, using a collaborative filtering algorithm to build a model using data from multiple users with similar preferences.

[1093] Content recommendations

[1094] The trained model is used to predict user preference patterns and recommend the most suitable content, thereby providing personalized content to users.

[1095] Complementing reviews

[1096] The server collects word-of-mouth information from other users and adds it to the recommendation list, making it easier for users to refer to the opinions and ratings of other users.

[1097] Displaying the recommendation list

[1098] The recommendation list sent from the server is visually displayed on the user's device, allowing the user to select and view content of interest from the list.

[1099] Acquiring emotion data

[1100] The server uses the smartphone's camera and microphone to obtain emotion data from the user's facial expressions and voice, using a neural network model for emotion analysis.

[1101] Real-time recommendation adjustment

[1102] The server adjusts the recommendation list in real time based on the acquired emotional data. For example, if a user is watching a scary movie, the server will recognize that emotional data in real time and recommend less scary content next.

[1103] Specific examples

[1104] If a user feels scared while watching "Horror Movie C," the server will recognize that emotion in real time and recommend a less scary movie for the next viewing, taking into account word-of-mouth reviews from other users to suggest more appropriate content.

[1105] Prompt Sentence Examples

[1106] "If it is detected in the user's viewing history that they felt scared while watching 'Horror Movie C,' recommend movies that are less scary based on their viewing history and rating data. Also, take into account word-of-mouth information from other users."

[1107] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1108] Step 1:

[1109] User Data Collection

[1110] Input: User's viewing history, ratings, search history, facial expression and voice data from the smartphone camera and microphone

[1111] The server collects user behavior data from a database in real time or periodically, and also collects facial expression and voice data using the smartphone's camera and microphone.

[1112] Output: Collected user viewing history, rating data, search history, facial expression data, and voice data

[1113] Step 2:

[1114] Data normalization

[1115] Input: Collected viewing history, rating data, search history

[1116] The server normalizes the collected rating data to a scale of 0 to 1. Specifically, if the rating is out of 10, it converts it to a scale of 0 to 1.

[1117] Output: Normalized rating data

[1118] Step 3:

[1119] Training a machine learning model

[1120] Input: Normalized rating data, viewing history, search history

[1121] The server uses the normalized data to train a collaborative filtering algorithm, building a model using data from multiple users with similar preferences.

[1122] Output: A trained machine learning model

[1123] Step 4:

[1124] Content recommendations

[1125] Input: Trained machine learning model, current user preference data

[1126] The server uses the trained model to recommend optimal content based on the current user preferences.

[1127] Output: Recommendation list

[1128] Step 5:

[1129] Complementing reviews

[1130] Input: Recommendation list, word-of-mouth data from other users

[1131] The server adds word-of-mouth information from other users to the recommendation list, thereby providing other users' opinions and ratings about the recommended content.

[1132] Output: Recommendation list with reviews added

[1133] Step 6:

[1134] Displaying the recommendation list

[1135] Input: Recommendation list with reviews

[1136] The server visually provides a recommendation list to the user's device, allowing the user to select and view content of interest from the list.

[1137] Output: Recommendation list displayed on the user's device

[1138] Step 7:

[1139] Acquiring emotion data

[1140] Input: User's facial expression data, voice data

[1141] The server uses the smartphone's camera and microphone to analyze data obtained from the user's facial expressions and voice using an emotion recognition algorithm, and uses a neural network to estimate the user's emotions.

[1142] Output: Real-time emotion data

[1143] Step 8:

[1144] Real-time recommendation adjustment

[1145] Input: Real-time emotion data, recommendation list

[1146] The server adjusts the recommendation list based on real-time emotion data. For example, if a user feels scared, the server will reflect that emotion and recommend a less scary movie as the next content to watch.

[1147] Output: A refined list of recommendations

[1148] 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.

[1149] 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.

[1150] 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.

[1151] [Fourth embodiment]

[1152] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1153] 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.

[1154] 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).

[1155] 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.

[1156] 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.

[1157] 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).

[1158] 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. 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.

[1159] 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.

[1160] 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.

[1161] 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.

[1162] 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.

[1163] 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.

[1164] 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."

[1165] The present invention is a system for automatically recommending content based on a user's preferences, and an embodiment thereof will be described in detail below.

[1166] User Data Collection

[1167] The means of collection is for the server to collect data such as users' viewing history, ratings, search history, etc. in real time or periodically. This is done based on the content viewed by the user and the ratings given.

[1168] Example: A server automatically records viewing history data and receives a rating score each time a user watches a movie.

[1169] Data normalization

[1170] Normalization is the process of converting collected data into a consistent format, making data of different formats and scales uniform and easier for algorithms to use.

[1171] Example: The server converts data rated by users on different scales (e.g., 10-point or 5-point) into a scale from 0 to 1.

[1172] Training an AI model

[1173] Next, a training means is used to train the machine learning algorithm based on the collected data, thereby learning user preference patterns and improving prediction accuracy.

[1174] Example: A server trains a collaborative filtering algorithm using data from a group of users with similar tastes to build a content recommendation model tailored to each user's preferences.

[1175] Complementing content recommendations and word-of-mouth information

[1176] Using the trained model, the system proposes content based on user preferences through a recommendation tool. It also collects user reviews and adds them to the recommendation list, allowing users to easily refer to other users' opinions and ratings.

[1177] Example: The server recommends two movies, "Action Movie A" and "Drama Movie B," based on the user's preferences, and adds evaluation comments from other users.

[1178] Displaying the recommendation list

[1179] The display method is a software interface that visually displays the recommendation list on the user's device, allowing the user to easily find new content.

[1180] Example: When a user opens a smartphone app, a recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[1181] Gathering feedback

[1182] Finally, we use a means to collect feedback, and store user ratings and comments in a database, which can be reflected in future recommendations to improve accuracy.

[1183] Example: If a user rates a movie they watched and leaves a comment saying "the story was good," the server collects that information and uses it for the next recommendation.

[1184] This allows the system of the present invention to provide a richer content experience to the user.

[1185] The processing flow will be explained below.

[1186] Step 1:

[1187] The server collects real-time user behavioral data such as viewing history, ratings, and search history. Every time a user watches a new movie or adds a rating, the information is stored in a database.

[1188] Specific operation: When a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[1189] Step 2:

[1190] The server normalizes the collected data, which involves converting data measured at different scales into a unified format.

[1191] Specific behavior: The server normalizes the user's rating data from a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[1192] Step 3:

[1193] The server uses the normalized data to train machine learning algorithms, which learn user preference patterns and improve prediction accuracy.

[1194] How it works: The server trains a collaborative filtering algorithm using the viewing history and rating data of multiple users with similar tastes.

[1195] Step 4:

[1196] Using the trained model, the server automatically recommends content based on the user's preferences, which can range from general popularity to niche content.

[1197] Specific behavior: The server analyzes that the user likes action movies and adds newly released action movies to the recommendation list.

[1198] Step 5:

[1199] The server collects word-of-mouth information about niche content and adds it to a recommendation list. This word-of-mouth information includes ratings and comments from other users.

[1200] Specific behavior: The server collects reviews of the movie "Drama Movie B" in which other users have rated it as "a moving story" and adds it to the recommendation list.

[1201] Step 6:

[1202] The server transmits the generated recommendation list to the user terminal.

[1203] Specific operation: When a user opens the smartphone app, the server synchronizes with the local database and sends the latest recommendation list.

[1204] Step 7:

[1205] The device displays the received recommendation list to the user, who can then select and view content of interest from the list.

[1206] Specific operation: When a user opens a smartphone app, the device displays recommended movies and music on the home screen. The user can tap to watch.

[1207] Step 8:

[1208] Users can watch the recommended content and then add ratings and comments, which will be reflected in the next recommendations.

[1209] Specific behavior: A user watches the movie "Drama Movie B," rates it highly, and comments that "it had a touching ending." This information will be used to generate the next recommendation.

[1210] Step 9:

[1211] The server collects user feedback and stores it in a database, which is then used to generate recommendations next time.

[1212] What happens: The server stores the ratings and comments users leave for "Drama Movie B" in a database and uses them for the next algorithm training.

[1213] These processing steps enable highly accurate content recommendations based on the user's preferences.

[1214] Example 1

[1215] 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."

[1216] Current content recommendation systems often lack sufficient accuracy in recommending content based on user preferences. They also often fail to provide effective recommendations to new users. Furthermore, they lack a mechanism for utilizing word-of-mouth information or efficiently collecting user feedback, which means they are unable to respond to real-time changes in user preferences. This can result in a poor user experience and a deterioration in overall system performance.

[1217] 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.

[1218] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing the recommendation list with word-of-mouth information, means for displaying the recommendation list on the user's terminal, means for collecting the user's viewing history from the terminal and transmitting it to the server, means for converting the collected data into a consistent format, and means for collecting feedback data based on user ratings and storing it in a database. This allows for accurate understanding of user preferences and enables effective content recommendations even for new users. Furthermore, utilizing word-of-mouth information increases user satisfaction, and collecting feedback continuously improves the accuracy of the system and the user experience.

[1219] "User data" refers to information that indicates a user's behavior and preferences, such as the user's viewing history, ratings, and search history.

[1220] "Collection Methods" refers to the technologies and processes used to obtain and transmit User Data to a server.

[1221] "Normalization methods" refer to techniques and processes that convert collected data into a consistent format and make it easier for algorithms to use.

[1222] A "machine learning algorithm" refers to a computer program that learns patterns and rules from data to make predictions and classifications.

[1223] "Training" refers to the process of training a machine learning algorithm using collected data.

[1224] "Recommendation method" refers to the technology or process that uses a trained model to recommend content based on user preferences.

[1225] "Word-of-mouth information" refers to evaluation comments and reviews of content left by other users.

[1226] A "recommendation list" refers to a list of content recommended to a user.

[1227] "Means for displaying" refers to the technology or process for visually presenting the recommendation list to the user's terminal.

[1228] "User terminal" refers to the device a user uses to interact with the system, such as a smartphone, PC, or tablet.

[1229] "Viewing history" refers to information about content that a user has viewed in the past.

[1230] "Consistent format" refers to a data format that unifies data of different scales and formats and makes it easy for algorithms to use.

[1231] "Feedback data" refers to the rating scores and comments provided by users after viewing content.

[1232] "Database" refers to a system for systematically storing and managing collected user data and feedback data.

[1233] The present invention is a system that automatically recommends content based on a user's preferences, and details thereof will be described below.

[1234] System Overview

[1235] The system mainly consists of the following elements:

[1236] How user data is collected

[1237] A means of normalizing data

[1238] A means of training machine learning algorithms

[1239] Recommendation methods

[1240] A means to complement word-of-mouth information

[1241] How to display a recommendation list

[1242] A means of gathering feedback

[1243] User Data Collection

[1244] The server collects users' viewing history, ratings, and search history in real time or periodically. The data includes the ID, start time, end time, and rating of the content viewed by the user. This data is stored in the server's database and used for analysis.

[1245] Example: When a user watches the movie "Action Movie A," the viewing history is stored in a database on the server.

[1246] Data normalization

[1247] The server converts the collected data into a consistent format, converting ratings from different scales (e.g., 10-point or 5-point scales) into a range between 0 and 1, and standardizing data in other formats, making the data consistent and suitable for training algorithms.

[1248] Example: Ratings of "4 out of 5" and "8 out of 10" are converted to the format "0.8".

[1249] Training an AI model

[1250] The server uses the normalized data to train machine learning algorithms, using techniques such as collaborative filtering and deep learning to learn user preference patterns, resulting in the creation of highly accurate recommendation models.

[1251] Example: The server uses past viewing and rating data to learn each user's preference patterns.

[1252] Complementing content recommendations and word-of-mouth information

[1253] The server uses the trained model to recommend content based on the user's preferences, and adds word-of-mouth information collected from other users to the recommendation list, making it easier for users to refer to other users' ratings.

[1254] Example: Recommend the movies "Action Movie A" and "Drama Movie B" and add ratings and comments from other users.

[1255] Displaying the recommendation list

[1256] The user device (smartphone, PC, tablet, etc.) visually displays the recommendation list sent from the server. When the user opens the application, they can select and watch content of their interest from the recommendation list displayed on the home screen.

[1257] Example: "Action Movie A" and "Drama Movie B" are displayed on the home screen.

[1258] Gathering feedback

[1259] The server stores user feedback (ratings and comments) in a database, which is used to improve the accuracy of recommendations next time.

[1260] Example: If a user watches the movie "Drama Movie B" and then comments that "the story was good," that information is saved on the server and used for the next recommendation.

[1261] Example prompts to input to the generative AI model

[1262] "Please enter the title and rating of the movie you watched:

[1263] Movie Title: "Action Movie A"

[1264] Review comment: "The action scenes were powerful."

[1265] Use this information to generate an up-to-date list of recommendations for your users."

[1266] This system enables highly accurate and personalized content recommendations to users, improving the user experience.

[1267] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1268] Step 1: Collect user data

[1269] Input: Information about the content viewed by the user (ID, viewing start time, end time, rating score)

[1270] How it works: Every time a user watches content, their device automatically sends viewing data to the server, which then records the data in a database. If the user enters a rating after viewing, the rating data is also sent to the server.

[1271] Output: Viewing history and rating data stored in a database

[1272] Step 2: Normalize the data

[1273] Input: Raw data in the database (viewing history, rating data, search history)

[1274] Specific operation: The server performs normalization processing on the collected data. For example, it standardizes all data rated on different scales (10-point or 5-point) to a scale from 0 to 1.

[1275] Output: Normalized, consistent data

[1276] Step 3: Training the AI ​​model

[1277] Input: Normalized viewing history data and rating data

[1278] Specific operation: The server uses the normalized data to train machine learning algorithms (e.g., collaborative filtering and deep learning), extracting features from the data and analyzing user preference patterns to build a model.

[1279] Output: A trained, highly accurate recommendation model

[1280] Step 4: Recommend content and complement reviews

[1281] Input: Trained AI model and customer reviews

[1282] Specific operation: The server uses the trained model to generate a content list optimized for each user. At the same time, it retrieves other users' evaluation comments from a word-of-mouth information database and adds them to the recommendation list.

[1283] Output: Recommendation list supplemented with reviews

[1284] Step 5: View the recommendation list

[1285] Input: Recommendation list sent from the server

[1286] Specific operation: When a user opens the application on a device (such as a smartphone or PC), a recommendation list sent from the server is displayed. The user can select and watch content that interests them from this list.

[1287] Output: Recommendation list displayed on the user's device

[1288] Step 6: Gather feedback

[1289] Input: Rating score and comments entered by users after viewing the content

[1290] How it works: After watching a video, users can leave a rating or comment, and that information is sent from the device to the server, which then stores this feedback in a database and uses it to train the recommendation model next time.

[1291] Output: User feedback data stored in a database

[1292] Through the above processing steps, the system can provide highly accurate content based on the user's preferences.

[1293] (Application example 1)

[1294] 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."

[1295] In today's world, it is extremely difficult for users to find content that suits their preferences from the vast amount of content available. Furthermore, conventional content recommendation systems are inadequately responsive to the specific devices and platforms of individual users. Furthermore, few systems offer a function that allows users to easily refer to word-of-mouth information, leaving users with a lack of reference information when making their selections.

[1296] 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.

[1297] In this invention, the server includes means for collecting user data, means for normalizing the data, means for training with a machine learning algorithm, means for recommending content based on user preferences using a smartphone, and means for supplementing word-of-mouth information. This allows users to easily find content that matches their preferences on their smartphone, and by referring to word-of-mouth information from other users, they can select content that will provide them with greater satisfaction.

[1298] "User data" refers to information such as a user's viewing history, ratings, and search history, and is data that indicates the user's preferences and behavior.

[1299] "Normalization" is the process of converting collected data into a consistent format and unifying data of different scales and formats.

[1300] A "machine learning algorithm" is a technology that analyzes data, learns user behavior patterns and preferences, and makes predictions and classifications.

[1301] "Recommendation" refers to selecting suitable content based on a user's preferences and presenting it to the user.

[1302] "Word-of-mouth information" refers to information such as ratings and comments provided by other users, and is opinions and impressions that can be used as a reference for making a selection.

[1303] A "user terminal" is an electronic device used by a user to display information and perform operations.

[1304] "Feedback" refers to reaction data such as ratings and comments provided by users after use.

[1305] "Collaborative filtering" is a method for recommending optimal content to individual users by comparing their preferences with data from other users who have similar preferences.

[1306] A "smartphone" is a portable information terminal with multiple functions, including communication, information processing, and application execution.

[1307] A "personalized movie recommendation app" is an application that recommends movies and TV dramas based on the preferences of each individual user.

[1308] The present invention is a personalized recommendation system that enables users to efficiently find content such as movies and dramas on their smartphones. The system of the present invention is configured as follows.

[1309] User Data Collection

[1310] The server collects data such as users' viewing history, ratings, and search history in real time or periodically. This provides basic data for understanding users' preferences. Specifically, each time a user watches a movie or drama, the server records the user's viewing history and rating score. Search history is also collected to understand the genres and titles in which the user is interested.

[1311] Data normalization

[1312] The server normalizes the collected data. User ratings are often rated on different scales (e.g., 10-point scale or 5-point scale). This data is converted to a unified scale (ranging from 0 to 1) to make it easier for machine learning algorithms to use.

[1313] Training an AI model

[1314] Using the collected and normalized data, the server trains a machine learning algorithm. This uses a collaborative filtering algorithm to build a content recommendation model for each user based on data from users with similar tastes. The trained model learns the user's preference patterns and can make more accurate recommendations.

[1315] Complementing content recommendations and word-of-mouth information

[1316] Using the trained model, the server recommends movies and TV shows based on the user's preferences to the smartphone. It also collects word-of-mouth information from other users and adds it to the recommendation list, allowing users to refer to other users' opinions and ratings.

[1317] Displaying the recommendation list

[1318] When a user opens the smartphone app, the recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[1319] Gathering feedback

[1320] When users leave ratings and comments on movies and TV shows they have watched, the server collects that feedback and stores it in a database. This information is then reflected in the next recommendation, further improving the system's prediction accuracy.

[1321] Through the above process, the system can provide the user with a personalized content experience. As a specific example, the following prompt sentence can be used:

[1322] Example prompt sentence:

[1323] Based on user ID 1's recent viewing history, what movies would you recommend to him?

[1324] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1325] Step 1: Collect user data

[1326] The server collects data such as the user's viewing history, rating scores, and search history. This data is collected in real time or periodically. For example, every time a user watches a movie, the viewing history and rating score are sent to the server and recorded in a database. The input is the user's viewing history, rating scores, and search history, and the output is the consistently recorded user data.

[1327] Step 2: Normalize the data

[1328] The server normalizes the collected user data into a consistent format. Specifically, it converts data rated on different scales into a scale from 0 to 1. This process unifies the data and makes it easier to use in machine learning algorithms. The input is the rating data on different scales, and the output is data normalized from 0 to 1.

[1329] Step 3: Training the AI ​​model

[1330] The server uses the normalized data to train a machine learning algorithm, which uses a collaborative filtering algorithm to build a content recommendation model based on data from a group of users with similar preferences. The input is the normalized user data, and the output is the trained recommendation model.

[1331] Step 4: Recommend content and complement reviews

[1332] The server uses the trained model to recommend content based on the user's preferences. In addition, word-of-mouth information collected from other users is added to the recommendation list, allowing users to refer to other users' ratings and comments. The input is the trained recommendation model and word-of-mouth information, and the output is a recommendation list with the word-of-mouth information added.

[1333] Step 5: View the recommendation list

[1334] When a user opens the smartphone app, the server sends a recommendation list to the user's device and displays it on the home screen. The user can select and watch content of interest from this list. The input is a recommendation list with added word-of-mouth information, and the output is a recommendation list displayed on the user's smartphone.

[1335] Step 6: Gather feedback

[1336] When users leave ratings or comments on movies or TV shows they have watched, the server collects that feedback and stores it in a database. The feedback data is reflected in the next recommendation. The input is the user's ratings and comments, and the output is the feedback data stored in the database.

[1337] The above is a specific processing flow in the system of the present invention.

[1338] 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.

[1339] The present invention is a system for automatically recommending content based on a user's preferences and emotions, and an embodiment thereof will be described in detail below.

[1340] User Data Collection

[1341] The means of collection is for the server to collect behavioral data such as user viewing history, ratings, and search history in real time or periodically, based on the content viewed by the user and the ratings given.

[1342] Example: If a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[1343] Data normalization

[1344] Normalization is the process of converting collected data into a consistent format, making data of different formats and scales uniform and easier for algorithms to use.

[1345] Example: The server normalizes the data that the user has rated on a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[1346] Training an AI model

[1347] Next, a training means is used to train the machine learning algorithm based on the collected data, thereby learning user preference patterns and improving prediction accuracy.

[1348] Example: A server trains a collaborative filtering algorithm using viewing history and rating data from multiple users with similar tastes.

[1349] Complementing content recommendations and word-of-mouth information

[1350] Using the trained model, the system proposes content based on user preferences through a recommendation tool. It also collects user reviews and adds them to the recommendation list, allowing users to easily refer to other users' opinions and ratings.

[1351] Example: The server recommends two movies, "Action Movie A" and "Drama Movie B," based on the user's preferences, and adds evaluation comments from other users.

[1352] Combining Emotion Engines

[1353] Furthermore, it combines an emotion engine to collect user emotional data, which recognizes emotions from users' facial expressions, voice, and text messages, and optimizes the recommendation list based on that data.

[1354] Example: The server monitors the user's facial expressions and tone of voice while watching the movie "Horror Movie C." If the server determines that the user is feeling scared, it will select a less frightening movie for the next recommendation.

[1355] Displaying the recommendation list

[1356] The display method is a software interface that visually displays the recommendation list on the user's device, allowing the user to easily find new content.

[1357] Example: When a user opens a smartphone app, a recommendation list sent from the server is displayed on the home screen, allowing the user to select and watch content of interest from the list.

[1358] Gathering feedback

[1359] Finally, we use a feedback collection method to store user ratings and comments in a database, which can be used to improve the accuracy of future recommendations.

[1360] Example: If a user rates a movie they watched and leaves a comment saying "the story was good," the server collects that information and uses it when generating the next recommendation.

[1361] In this way, the system of the present invention can realize highly accurate and personalized content recommendations based on the user's preferences and emotions, and provide the user with a richer content experience.

[1362] The processing flow will be explained below.

[1363] Step 1:

[1364] The server collects real-time user behavioral data such as viewing history, ratings, and search history. Every time a user watches a new movie or adds a rating, the information is stored in a database.

[1365] Specific operation: When a user watches the movie "Action Movie A" and rates it 4 out of 5, the server automatically collects the user's viewing history and rating and adds them to the database.

[1366] Step 2:

[1367] The server normalizes the collected data, which involves converting data measured at different scales into a unified format.

[1368] Specific behavior: The server normalizes the user's rating data from a 10-point scale to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[1369] Step 3:

[1370] The server uses the normalized data to train machine learning algorithms, which learn user preference patterns and improve prediction accuracy.

[1371] How it works: The server trains a collaborative filtering algorithm using the viewing history and rating data of multiple users with similar tastes.

[1372] Step 4:

[1373] Using the trained model, the server automatically recommends content based on the user's preferences, which can range from general popularity to niche content.

[1374] Specific behavior: The server analyzes that the user likes action movies and adds newly released action movies to the recommendation list.

[1375] Step 5:

[1376] The server collects word-of-mouth information about niche content and adds it to a recommendation list. This word-of-mouth information includes ratings and comments from other users.

[1377] Specific behavior: The server collects reviews of the movie "Drama Movie B" in which other users have rated it as "a moving story" and adds it to the recommendation list.

[1378] Step 6:

[1379] The server collects the user's emotion data using an emotion engine that recognizes the user's emotions, which includes the process of recognizing emotions from the user's facial expressions, voice, and text messages.

[1380] Specific operation: The server analyzes the user's facial expressions and tone of voice while watching the movie "Horror Movie C," and if it determines that the user is feeling scared, it stores that data.

[1381] Step 7:

[1382] The server optimizes the recommendation list based on the emotional data, adjusting the recommendations of relevant content according to the user's emotional state.

[1383] Specific behavior: The server recognizes that the user is prone to fear and generates a recommendation list that reduces the proportion of horror movies and increases the proportion of comedy movies.

[1384] Step 8:

[1385] The server transmits the generated recommendation list to the user terminal.

[1386] Specific operation: When a user opens the smartphone app, the server synchronizes with the local database and sends the latest recommendation list.

[1387] Step 9:

[1388] The terminal displays the received recommendation list to the user, who can then select and view content of interest from the list.

[1389] Specific operation: When a user opens a smartphone app, the device displays recommended movies and music on the home screen. The user can tap to watch.

[1390] Step 10:

[1391] Users can watch the recommended content and then add ratings and comments, which will be reflected in the next recommendations.

[1392] Specific behavior: A user watches the movie "Drama Movie B," rates it highly, and comments that "it had a touching ending." This information will be used to generate the next recommendation.

[1393] Step 11:

[1394] The server collects user feedback and stores it in a database, which is then used to generate recommendations next time.

[1395] What happens: The server stores the ratings and comments users leave for "Drama Movie B" in a database and uses them for the next algorithm training.

[1396] These processing steps enable highly accurate content recommendations based on the user's preferences and emotions.

[1397] Example 2

[1398] 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."

[1399] Conventional content recommendation systems rely on recommendation functions based on users' viewing history and rating data, but do not take into account the user's emotional state. This makes it difficult to provide optimal content to individual users, and there is a need to improve the user experience.

[1400] 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.

[1401] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing the recommendation list with word-of-mouth information, means for collecting emotional data to optimize the recommendation list, means for displaying the recommendation list on a user terminal, and means for collecting feedback, thereby enabling highly personalized recommendations that take into account not only the preferences but also the emotional state of each individual user.

[1402] "Means for collecting user data" refers to a function for collecting behavioral data such as user viewing history, ratings, and search history in real time or periodically.

[1403] "Means for normalizing collected data" refers to the ability to perform a process to convert data of different formats or scales into a consistent format.

[1404] The "means for training a machine learning algorithm" is a function for training a machine learning algorithm using normalized data to improve the accuracy of the model.

[1405] "Means for recommending content based on user preferences" is a function that uses a trained model to suggest content that best suits the user's preferences.

[1406] "Means for supplementing recommendation lists with word-of-mouth information" refers to a function that collects ratings and comments from other users and adds them to the recommendation list.

[1407] "Means of collecting emotional data to optimize recommendation lists" refers to a function that recognizes emotions from the user's facial expressions, voice, and text messages, and uses that data to adjust the recommendation list to be more optimal for the individual.

[1408] The "means for displaying the recommendation list on the user terminal" is a function for visually displaying the recommendation list sent from the server on the interface of the user terminal.

[1409] "Means for collecting feedback" is a function that collects ratings and comments from users, stores that information in a database, and reflects it in the next recommendation.

[1410] "Collaborative filtering" is a machine learning algorithm that recommends appropriate content to individual users based on the viewing history and rating data of multiple users with similar tastes.

[1411] This invention is a system that automatically recommends content based on user preferences and emotions. This system mainly consists of a server and a user terminal. The server is responsible for data collection, data normalization, machine learning algorithm training, emotion data collection, recommendation generation, and feedback collection. The user terminal provides an interface with the user and has the function of displaying a recommendation list.

[1412] User Data Collection

[1413] The server collects user behavioral data such as viewing history, ratings, and search history in real time or periodically. This allows the server to accurately grasp user preferences. For example, if a user watches "Action Movie A" and rates it 4 out of 5, the viewing history and rating are added to the database.

[1414] Data normalization

[1415] The server then processes the collected data to convert it into a consistent format, unifying data from different formats and scales, making it easier for machine learning algorithms to use. For example, data rated by a user on a 10-point scale is normalized to a 0-1 scale. For example, an 8-point rating is stored as 0.8.

[1416] Training an AI model

[1417] The server trains machine learning algorithms based on the normalized data, which allows it to learn user preference patterns and improve prediction accuracy. For example, a collaborative filtering algorithm is used to learn the viewing history and rating data of multiple users with similar preferences.

[1418] Complementing content recommendations and word-of-mouth information

[1419] Using the trained model, the server recommends content based on the user's preferences. It also supplements the recommendation list with word-of-mouth information to provide other users' opinions and ratings. For example, it recommends "Action Movie A" and "Drama Movie B" to the user, along with other users' ratings and comments.

[1420] Combining Emotion Engines

[1421] The server uses an emotion engine to collect emotional data from the user's facial expressions, voice, and text messages, and optimizes the recommendation list. This process allows the server to provide content that best suits the user's current emotional state. For example, if the user expresses fear while watching "Horror Movie C," the server will select a less frightening movie for the next recommendation list.

[1422] Displaying the recommendation list

[1423] The user device uses a software interface that visually displays the recommendation list sent from the server, allowing the user to easily find new content. For example, when a user opens a smartphone app, the recommendation list is displayed on the home screen.

[1424] Gathering feedback

[1425] The server collects user ratings and comments in a database and reflects this information in the next recommendation. For example, if a user rates a movie they watched and leaves a comment saying "the story was good," that information will be used when generating the next recommendation.

[1426] In this way, the system of the present invention can realize highly accurate and personalized content recommendations based on the user's preferences and emotions, providing the user with a richer content experience.

[1427] Prompt Sentence Examples

[1428] "A user watched the movie 'Action Movie A' and rated it 4 out of 5. Please recommend new action movies based on their viewing history and ratings."

[1429] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1430] Step 1: Collect user data

[1431] The server collects user behavioral data, such as viewing history, ratings, and search history, in real time or periodically. Specifically, when a user watches a particular movie and enters a rating, the information is sent to the server. The input data is the movie title "Action Movie A" and the rating "4." The output is that these data are stored in a database.

[1432] Step 2: Normalize the data

[1433] The server then processes the collected raw user data to convert it into a consistent format. For example, it normalizes the data that the user has rated on a 10-point scale to a 0-1 scale. This unifies data from different formats and scales. The input data is a rating of 4, which is normalized to 0.8, and the normalized data is stored as the output.

[1434] Step 3: Training the AI ​​model

[1435] The server uses the normalized data to train a machine learning algorithm. Specifically, it uses a collaborative filtering algorithm to train the model on the viewing history and rating data of multiple users with similar preferences. The normalized rating data (0.8) and other users' data are used as input, and the output is the trained model.

[1436] Step 4: Recommend content and complement reviews

[1437] Using the trained model, the server recommends content based on the user's preferences. The server also supplements the recommendation list with reviews collected from other users. The input data is the trained model and review data, and the output is a recommendation list tailored to the user. Specifically, "Action Movie A" and "Drama Movie B" are included in the list, and reviews for each are displayed.

[1438] Step 5: Combining Emotion Engines

[1439] The server uses an emotion engine to collect user emotional data and optimize the recommendation list. It recognizes emotions from the user's facial expressions, voice, and text messages and makes adjustments based on that data. The input data is the user's emotional data while watching, and the output is a recommendation list that takes emotions into consideration. Specifically, if the server recognizes that the user is feeling scared, the next movie recommendation will be a less scary one.

[1440] Step 6: Displaying the recommendation list

[1441] The device uses a software interface that visually presents the recommendation list sent from the server. The input is the recommendation list data from the server, and the output is the list displayed on the device screen. When a user opens the smartphone app, a list of related recommended movies and TV shows is displayed on the home screen.

[1442] Step 7: Gather feedback

[1443] The server collects ratings and comments from users and reflects this information in the next recommendation. The input data are user ratings and comments, which are saved in a database and used to generate the next recommendation. The output is an updated database. For example, if a user rates a story as "good," that information is reflected in the server.

[1444] (Application example 2)

[1445] 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."

[1446] Conventional content recommendation systems make recommendations based on user preferences and behavioral data, but do not take into account the user's real-time emotions when recommending content. As a result, content that does not match the user's momentary emotions or state may be recommended, resulting in a poor user experience. The present invention aims to provide a system that combines the user's real-time emotional information to realize highly accurate and personalized content recommendations.

[1447] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1448] In this invention, the server includes means for collecting user data, means for normalizing the collected data, means for training a machine learning algorithm using the normalized data, means for recommending content based on user preferences using the trained model, means for supplementing a recommendation list with word-of-mouth information, means for displaying the recommendation list on a user terminal, means for acquiring emotion data using the user's facial expressions and voice, and means for adjusting the recommendation list in real time based on the emotion data. This makes it possible to recommend more appropriate content based on the user's real-time emotions.

[1449] "User data" refers to information about a user's behavior and preferences, such as their viewing history, ratings, and search history.

[1450] "Means for collecting" refers to a method or device by which the server collects behavioral data such as user viewing history and ratings in real time or periodically.

[1451] "Normalization means" refers to a method or device that converts collected data into a consistent format and unifies data of different formats or scales.

[1452] "Means for training" refers to a method or apparatus that uses normalized data to train a machine learning algorithm.

[1453] "Means for recommending" refers to a method or apparatus that uses a trained model to suggest content based on a user's preferences.

[1454] The "means for supplementing word-of-mouth information" refers to a method or device for collecting opinions and ratings from other users and adding them to a recommendation list.

[1455] The "means for displaying" refers to a software interface that visually presents the recommendation list on the user terminal.

[1456] "Means for acquiring emotion data using facial expressions and voice" refers to a method or device for capturing a user's facial expressions and voice in real time and recognizing their emotions.

[1457] "Means for adjusting a recommendation list in real time" refers to a method or apparatus for optimizing a recommendation list in real time based on acquired emotion data.

[1458] "Machine learning algorithm" refers to an algorithm that learns user preference patterns based on collected data and improves prediction accuracy.

[1459] "Collaborative filtering" refers to a technique for improving recommendation accuracy by using data from multiple users with similar preferences.

[1460] The present invention is a system for automatically recommending content based on a user's preferences and emotions, and an embodiment thereof will be described in detail below.

[1461] System configuration

[1462] The system includes means for collecting user data, normalizing the data, training machine learning algorithms, recommending content, complementing reviews, displaying recommendation lists, acquiring sentiment data, and adjusting recommendations in real time.

[1463] Data collection

[1464] The server collects user behavioral data, such as viewing history, ratings, and search history, in real time or periodically. To do so, it obtains the user's viewing history and ratings from a database. It also collects facial expressions and voice data using the smartphone's camera and microphone.

[1465] Data normalization

[1466] The server converts the collected data into a consistent format, specifically normalizing the rating data to a scale of 0 to 1. This unifies the variations in data format and scale, making it easier for algorithms to use.

[1467] Training machine learning algorithms

[1468] The server uses the normalized data to train a machine learning algorithm, using a collaborative filtering algorithm to build a model using data from multiple users with similar preferences.

[1469] Content recommendations

[1470] The trained model is used to predict user preference patterns and recommend the most suitable content, thereby providing personalized content to users.

[1471] Complementing reviews

[1472] The server collects word-of-mouth information from other users and adds it to the recommendation list, making it easier for users to refer to the opinions and ratings of other users.

[1473] Displaying the recommendation list

[1474] The recommendation list sent from the server is visually displayed on the user's device, allowing the user to select and view content of interest from the list.

[1475] Acquiring emotion data

[1476] The server uses the smartphone's camera and microphone to obtain emotion data from the user's facial expressions and voice, using a neural network model for emotion analysis.

[1477] Real-time recommendation adjustment

[1478] The server adjusts the recommendation list in real time based on the acquired emotional data. For example, if a user is watching a scary movie, the server will recognize that emotional data in real time and recommend less scary content next.

[1479] Specific examples

[1480] If a user feels scared while watching "Horror Movie C," the server will recognize that emotion in real time and recommend a less scary movie for the next viewing, taking into account word-of-mouth reviews from other users to suggest more appropriate content.

[1481] Prompt Sentence Examples

[1482] "If it is detected in the user's viewing history that they felt scared while watching 'Horror Movie C,' recommend movies that are less scary based on their viewing history and rating data. Also, take into account word-of-mouth information from other users."

[1483] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1484] Step 1:

[1485] User Data Collection

[1486] Input: User's viewing history, ratings, search history, facial expression and voice data from the smartphone camera and microphone

[1487] The server collects user behavior data from a database in real time or periodically, and also collects facial expression and voice data using the smartphone's camera and microphone.

[1488] Output: Collected user viewing history, rating data, search history, facial expression data, and voice data

[1489] Step 2:

[1490] Data normalization

[1491] Input: Collected viewing history, rating data, search history

[1492] The server normalizes the collected rating data to a scale of 0 to 1. Specifically, if the rating is out of 10, it converts it to a scale of 0 to 1.

[1493] Output: Normalized rating data

[1494] Step 3:

[1495] Training a machine learning model

[1496] Input: Normalized rating data, viewing history, search history

[1497] The server uses the normalized data to train a collaborative filtering algorithm, building a model using data from multiple users with similar preferences.

[1498] Output: A trained machine learning model

[1499] Step 4:

[1500] Content recommendations

[1501] Input: Trained machine learning model, current user preference data

[1502] The server uses the trained model to recommend optimal content based on the current user preferences.

[1503] Output: Recommendation list

[1504] Step 5:

[1505] Complementing reviews

[1506] Input: Recommendation list, word-of-mouth data from other users

[1507] The server adds word-of-mouth information from other users to the recommendation list, thereby providing other users' opinions and ratings about the recommended content.

[1508] Output: Recommendation list with reviews added

[1509] Step 6:

[1510] Displaying the recommendation list

[1511] Input: Recommendation list with reviews

[1512] The server visually provides a recommendation list to the user's device, allowing the user to select and view content of interest from the list.

[1513] Output: Recommendation list displayed on the user's device

[1514] Step 7:

[1515] Acquiring emotion data

[1516] Input: User's facial expression data, voice data

[1517] The server uses the smartphone's camera and microphone to analyze data obtained from the user's facial expressions and voice using an emotion recognition algorithm, and uses a neural network to estimate the user's emotions.

[1518] Output: Real-time emotion data

[1519] Step 8:

[1520] Real-time recommendation adjustment

[1521] Input: Real-time emotion data, recommendation list

[1522] The server adjusts the recommendation list based on real-time emotion data. For example, if a user feels scared, the server will reflect that emotion and recommend a less scary movie as the next content to watch.

[1523] Output: A refined list of recommendations

[1524] 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.

[1525] 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.

[1526] 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 robot 414.

[1527] 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.

[1528] 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.

[1529] 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.

[1530] 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).

[1531] 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.

[1532] 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."

[1533] 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.

[1534] 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).

[1535] 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.

[1536] 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.

[1537] 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.

[1538] 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.

[1539] 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.

[1540] 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.

[1541] 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.

[1542] 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.

[1543] 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.

[1544] 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.

[1545] The following is further disclosed regarding the above embodiment.

[1546] (Claim 1)

[1547] a means for collecting user data;

[1548] a means of normalizing the collected data;

[1549] means for training a machine learning algorithm using the normalized data;

[1550] a means for recommending content based on user preferences using the trained model;

[1551] A means to supplement the recommendation list with word-of-mouth information,

[1552] a means for displaying the recommendation list on a user terminal;

[1553] A system including:

[1554] (Claim 2)

[1555] 10. The system of claim 1, further comprising means for collecting feedback from users.

[1556] (Claim 3)

[1557] The system of claim 1 , wherein the machine learning algorithm uses collaborative filtering.

[1558] "Example 1"

[1559] (Claim 1)

[1560] a means for collecting user data;

[1561] a means of normalizing the collected data;

[1562] means for training a machine learning algorithm using the normalized data;

[1563] a means for recommending content based on user preferences using the trained model;

[1564] A means to supplement the recommendation list with word-of-mouth information,

[1565] a means for displaying the recommendation list on a user terminal;

[1566] A means for collecting a user's viewing history from the terminal and transmitting the history to a server;

[1567] means for converting the data collected by said server into a consistent format;

[1568] a means for collecting and storing user-rated feedback data in a database;

[1569] A system including:

[1570] (Claim 2)

[1571] 10. The system of claim 1, further comprising means for collecting feedback from users.

[1572] (Claim 3)

[1573] The system of claim 1 , wherein the machine learning algorithm uses collaborative filtering.

[1574] "Application Example 1"

[1575] (Claim 1)

[1576] a means for collecting user data;

[1577] a means of normalizing the collected data;

[1578] means for training a machine learning algorithm using the normalized data;

[1579] a means for recommending content based on user preferences using the trained model;

[1580] A means to supplement the recommendation list with word-of-mouth information,

[1581] a means for displaying the recommendation list on a user terminal;

[1582] The user terminal is a smartphone;

[1583] A system including:

[1584] (Claim 2)

[1585] 10. The system of claim 1, further comprising means for collecting feedback from users.

[1586] (Claim 3)

[1587] 10. The system of claim 1, wherein the machine learning algorithm uses collaborative filtering and the system is a personalized movie recommendation app.

[1588] "Example 2: Combining Emotion Engines"

[1589] (Claim 1)

[1590] a means for collecting user data;

[1591] a means of normalizing the collected data;

[1592] means for training a machine learning algorithm using the normalized data;

[1593] a means for recommending content based on user preferences using the trained model;

[1594] A means to supplement the recommendation list with word-of-mouth information,

[1595] A means of collecting sentiment data to optimize the recommendation list;

[1596] a means for displaying the recommendation list on a user terminal;

[1597] a means of collecting feedback;

[1598] A system including:

[1599] (Claim 2)

[1600] 10. The system of claim 1, further comprising means for collecting feedback from users.

[1601] (Claim 3)

[1602] The system of claim 1 , wherein the machine learning algorithm uses collaborative filtering.

[1603] "Application example 2 when combining emotion engines"

[1604] New Claims

[1605] (Claim 1)

[1606] a means for collecting user data;

[1607] a means of normalizing the collected data;

[1608] means for training a machine learning algorithm using the normalized data;

[1609] a means for recommending content based on user preferences using the trained model;

[1610] A means to supplement the recommendation list with word-of-mouth information,

[1611] a means for displaying the recommendation list on a user terminal;

[1612] A means for acquiring emotion data using a user's facial expression and voice;

[1613] means for adjusting a recommendation list in real time based on the emotion data;

[1614] A system including:

[1615] (Claim 2)

[1616] 10. The system of claim 1, further comprising means for collecting feedback from users.

[1617] (Claim 3)

[1618] The system of claim 1 , wherein the machine learning algorithm uses collaborative filtering. [Explanation of symbols]

[1619] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. a means for collecting user data; a means of normalizing the collected data; means for training a machine learning algorithm using the normalized data; a means for recommending content based on user preferences using the trained model; A means to supplement the recommendation list with word-of-mouth information, a means for displaying the recommendation list on a user terminal; A system including:

2. 10. The system of claim 1, further comprising means for collecting feedback from users.

3. The system of claim 1 , wherein the machine learning algorithm uses collaborative filtering.

Images