Content suggestion method

Abstract

To provide a content proposal method for enabling a user to browse content suitable for the user's feeling.SOLUTION: A content proposal method includes a step of receiving user information including emotion information of a user from a user terminal, a step of converting a data format of the user information into user feature information that can be compared with content feature information, and a step of referring to content data in which one or a plurality of contents suitable for the emotion of the user are browsable from the content data, and generating proposal information in which one or a plurality of contents stored in association with the content feature information are extracted on the basis of similarity between the user feature information and each content feature information.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a content suggestion method that utilizes emotion analysis data to determine and extract content that is suited to a user's emotions and makes the content viewable by the user. [Background technology]

[0002] When searching for books at bookstores, libraries, e-book subscription services, online stores, etc., keywords and book categories entered by the user are used. By entering keywords or selecting one or more categories from a hierarchical structure, users can narrow down the vast collection of books to a group of books similar to the desired book. This allows users to select books of interest from that group of books.

[0003] Patent Document 1 discloses a library book search and book management support system for assisting book searches. This library book search and book management support system has a book storage means for storing book information to which one or more classification tags are assigned for classifying books in the library. Multiple classification tags are set for each managed book, and are assigned hierarchically from major classifications to minor classifications. A user can select one or more classification tags on the terminal they use. The library book search and book management support system can extract a group of books based on one or more classification tags entered by the user. This allows a user to view a group of books assigned with classification tags they have selected and select a desired book from that group of books. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2023-098146 Summary of the Invention [Problem to be solved by the invention]

[0005] However, when searching by keyword, the keywords associated with the book and the keywords entered by the user must match exactly, so when the user enters arbitrary keywords, the search is likely to be missed, while when trying to prevent this by entering simple keywords, the search results are not narrowed down sufficiently, which is a hassle.Even when searching by category, users must select from a large number of hierarchical categories, making it difficult for them to find the book they want.

[0006] Furthermore, for a user to enjoy reading, it is preferable that the desired book matches a book that is appropriate for the user's emotions. However, a user's self-perceived emotions are not always accurate, and if the user selects a book without objectively evaluating their own emotions, they may end up choosing a book that is inappropriate for their emotions. In this regard, conventional book search methods require the user to input keywords, categories, etc., but it is not easy for the user to objectively view their own emotions and select books that are appropriate for those emotions. The books narrowed down by the user are not necessarily appropriate for the user's emotions. If the book selected by the user is not appropriate for the user's emotions, not only will the user be unable to read the book thoroughly, but the user may also develop a negative image of reading. Therefore, conventional book search methods require assistance in selecting books that are appropriate for the user's emotions.

[0007] In particular, while children attending schools and other educational institutions may have time to read books of their own choosing, it is difficult for children who are still developing to find books that suit their emotions at the time or to encounter books that match those emotions.As a result, if a child chooses a book that does not suit their emotions, there is a possibility that the child will lose their motivation to actively read and that the entrance to new learning that is tailored to their rich emotions will be narrowed.For this reason, even in such educational institutions, it is necessary to support the selection of books that suit the emotions of children and students.

[0008] Furthermore, while there are a variety of content available in the world, including not only books but also movies, events, web content, and games, it has been difficult for users to find content that matches their real-time emotions when searching for content from the vast selection of content. [Means for solving the problem]

[0009] The present invention has been made in view of the above-mentioned problems, and an object of the present invention is to provide a content suggestion method that enables a user to view content that suits the user's emotions.

[0010] In other words, the present invention is a content suggestion method that refers to content data that associates each of a plurality of pieces of content with content feature information generated by converting the data format of each piece of content, and enables a user to view one or more pieces of content that suit the user's emotions from the content data, and includes the steps of receiving user information including at least the user's emotional information from the user terminal of the user, converting the data format of the received user information into user feature information that can be compared with the content feature information, referring to the content data, and generating suggested information that extracts one or more pieces of content stored in association with the content feature information based on the similarity between the user feature information and each piece of content feature information, and transmitting suggested data display information to the user terminal for displaying the suggested information on the user terminal. [Effects of the Invention]

[0011] According to the content suggestion method of the present invention, the data format of user information including user emotion information can be converted into user characteristic information. The content suggestion method can generate suggested information by extracting one or more pieces of content stored in association with the content characteristic information based on the similarity between the user characteristic information and a plurality of pieces of content stored in association with the content characteristic information. The content suggestion method can transmit suggested data display information to the user terminal for displaying the suggested information on the display screen of the user terminal. Therefore, the content suggestion method can enable the user to view one or more pieces of content appropriate to the user's emotion on the screen of the user terminal when searching for content. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a diagram showing an overview of a content suggestion system according to an embodiment of the present invention. [Figure 2] 1 is a diagram showing a configuration of a content proposal system according to an embodiment of the present invention. [Figure 3] FIG. 1 is a diagram showing a flow of a content suggestion method according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0013] The content suggestion method of the present invention will be described in detail below with reference to the accompanying drawings.

[0014] FIG. 1 is a diagram showing an outline of a content proposal system 100 capable of executing a content proposal method according to one embodiment of the present invention. The content proposal system 100 allows a user to browse content that suits the user's emotions. The content may be various, whether analog or digital, and includes books, movies, events, web content, games, etc. In this embodiment, an e-book will be used as an example of the content.

[0015] The content proposal system 100 can communicate with a user terminal 200 used by a user, an analysis server 300 for analyzing emotions from voice data, and a conversion server 400 for generating numerical vectors from digital data via a computer network such as the Internet. Examples of the user terminal 200 include a smartphone, a tablet, a PC, and a wearable device. The content proposal system 100 has a processing unit 101 such as a CPU, a communication unit 102 such as a communication interface, and a storage unit 103 including a storage device or recording medium such as a memory or a hard disk, and can execute programs for performing each process described below. The content proposal system 100 may include one or more devices, computers, or servers. The programs can be stored in the storage unit 103 or a storage medium accessible from the content proposal system 100 and executed by the processing unit 101.

[0016] 2 is a diagram showing the configuration of a content proposal system 100 according to one embodiment of the present invention. The content proposal system 100 includes a content data storage unit 104, a user information receiving unit 105, an emotion information transmitting unit 106, an emotion information receiving unit 107, a user information transmitting unit 108, a user characteristic information receiving unit 109, a proposal information generating unit 110, and a proposal information transmitting unit 111. The content data storage unit 104 stores content data that associates each of a plurality of pieces of content with content characteristic information generated by converting the data format of each piece of content. The content data storage unit 104 of this embodiment stores e-books as the content.

[0017] The content feature information is a numerical vector generated by vectorizing the text data of the e-book, which is the content. The vectorization of an e-book involves converting the text data into a multidimensional real-valued vector using a technique such as a neural network. That is, the content feature information is obtained by vectorizing the text of the e-book using embedding, a type of natural language processing, and is an ordered array of real values. In addition to embedding, various other vectorization techniques are available, such as vectorization based on the frequency of occurrence of words in a sentence. Furthermore, the method of generating the content feature information is not limited to the method of converting into a numerical vector using embedding or the like; various methods can be used to convert the content feature information into a data format that can be compared with the user feature information described below.

[0018] The content data is a set of multiple e-books associated with content feature information generated by vectorizing each e-book. In this embodiment, the text of the e-books is vectorized, but the title or synopsis of the e-books may also be vectorized. Furthermore, the content is not limited to text data, and images, videos, audio, etc. may also be used.

[0019] 3 shows the flow of the content proposal method executed by the content proposal system 100. The content proposal system 100 can receive user information including at least emotional information of the user from the user terminal 200 used by the user via the user information receiving unit 105 (S301). The user information is input by the user operating a user information input screen displayed by a web browser accessible from the user terminal 200 or an application installed in the user terminal 200. The user information input screen allows the user to input personal attributes indicating the user's own characteristics, such as the user's gender and age, multiple selection items, arbitrary text, and the user's emotional information.

[0020] The plurality of selection items are set as options related to detailed information about the user, and the user can select one or more selection items on the user information input screen. Examples of the selection items include items that indicate the user's mood, such as wanting to be excited or moved, and items that indicate the user's skills, such as knowledge and qualifications, as long as they are items that serve as judgment factors for matching with content suitable for the user. Furthermore, while the user information input screen allows any text to be input, it may also be configured such that a question is set in advance and an answer to the question can be input as text.

[0021] The emotional information that can be input on the user information input screen includes not only emotional information that can be directly input using text, icons, etc., but also digital data such as facial images and voices that can be analyzed to read emotions. In this embodiment, voice based on the user's vocalization can be input as the emotional information on the user information input screen. The user terminal 200 acquires voice based on the user's vocalization using the user terminal 200 from a microphone built into the user terminal 200, and converts the voice into digital voice data. The voice data as the emotional information is analyzed in an analysis server 300 (described later), and the content proposal system 100 can recognize the user's emotion.

[0022] The content proposal system 100 can transmit voice data as the emotional information included in the user information to the analysis server via the emotional information transmission unit 106 (S302). The voice data transmitted to the analysis server by the emotional information transmission unit 106 is voice data received from the user terminal 200 converted into a data format suitable for analysis by the analysis server 300. In this embodiment, the content proposal system 100 converts the voice data into a data format suitable for analysis by the analysis server 300, but it is also acceptable if the data format is converted by the user terminal 200 and the voice data, which has already undergone data format conversion, is received from the user terminal 200.

[0023] The content proposal system 100 can receive emotional information from the analysis server 300 via the emotional information receiving unit 107 (S303). The emotional information is converted by the analysis server from voice data based on the user's vocalization. The emotional information is composed of a plurality of items, and a value is set for each item. The emotional information items include sad, calm, anger, energy, and joy, and a value is set for each item. The types and number of emotional information items can be changed as appropriate.

[0024] Furthermore, while the analysis server 300 generates the emotion information using multiple items and their values, the configuration of the emotion information is not limited to this and may generate text data, etc. The content proposal system 100 can accept both emotion information directly input by the user and analyzable emotion information. The former is characterized by being able to acquire the user's subjective emotion information, while the latter is characterized by being able to acquire the user's objective emotion information regardless of the user's subjectivity. In this embodiment, the emotion information included in the user information is voice data, and therefore the emotion information is analyzed by the analysis server 300. However, if the emotion information included in the user information is directly input using text, icons, or the like, the content proposal system 100 does not necessarily need to include the emotion information transmitting unit 106 and the emotion information receiving unit 107.

[0025] In addition, in this embodiment, voice data as emotional information included in the user information is transmitted to an analysis server, and the analysis server generates emotional information, which is then received by the content proposal system 100. However, the function of the analysis server may be implemented in the content proposal system 100. In this case, the content proposal system 100 is configured to further include an emotion analysis unit for analyzing emotional information from the voice data.

[0026] The content proposal system 100 can convert the data format of the user information including emotion information received from the analysis server 300 into user feature information that can be compared with the content feature information (S304). The user feature information is obtained by vectorizing the user information through embedding, similar to the content feature information. The text data of the emotion information is vectorized through the embedding by inputting each item and its value included in the emotion information. In this embodiment, since the vectorization of the user information is performed by the conversion server 400, the content proposal system 100 transmits the user information to the conversion server 400 via the user information transmission unit 108 (S304). The conversion server 400 generates user feature information by vectorizing the user information. The user feature information is a numerical vector generated by vectorizing the text data of the user's personal attributes, such as gender and age, multiple selection items, arbitrary text, and the emotion information included in the user information. As described above, it is sufficient that the user information includes at least emotion information, and that the conversion server 400 is capable of converting the data format of the emotion information to generate user characteristic information.

[0027] The content proposal system 100 can receive user characteristic information generated from the user information by the user characteristic information receiving unit 109 (S305). Note that in this embodiment, the user information is converted into user characteristic information by the conversion server 400 and the user characteristic information is received from the conversion server 400. However, the function of the conversion server 400 may be implemented in the content proposal system 100. In this case, the content proposal system 100 further includes a user characteristic information generating unit for generating the user characteristic information from the user information, and the user characteristic information is generated by the user characteristic information generating unit in S304. Furthermore, the content proposal system 100 does not need to include the user information transmitting unit 108 and the user characteristic information receiving unit 109, and can skip S304 of transmitting the user information to the conversion server 400 and S305 of receiving the user characteristic information from the conversion server 400.

[0028] The content proposal system 100 can generate proposal information by using the proposal information generation unit 110, by referencing the content data stored in the content data storage unit 104, and extracting one or more pieces of content stored in association with the content feature information based on the similarity between the user feature information and each piece of content feature information (S306). Cosine similarity can be used as a method for evaluating the similarity between the user feature information, which is a numerical vector, and the content feature information. Cosine similarity indicates the similarity between vectors and is a method for determining the similarity between two vectors based on the cosine value of the angle between the two vectors. Cosine similarity is normalized within a range from -1, indicating dissimilarity, to +1, indicating similarity.

[0029] Therefore, the proposal information generation unit 110 calculates the cosine similarity between the user feature information and the content feature information, and determines that the two are similar if the cosine similarity exceeds a threshold, and determines that the two are not similar if the cosine similarity is equal to or less than the threshold. The threshold can be set to any value, such as 0.2. Note that the proposal information generation unit 110 calculates the evaluation of the similarity between the user feature information and the content feature information, which are numerical vectors, using the cosine similarity, but this is not limited to this, and various methods such as Euclidean distance can be applied as long as they can evaluate the similarity between two vectors.

[0030] The suggestion information generator 110 compares each piece of content feature information stored in the content data storage unit 104 with user feature information obtained by vectorizing the user information. The suggestion information generator 110 can generate suggestion information by extracting one or more e-books stored in association with content feature information for which a similarity exceeding a predetermined threshold has been calculated, and / or one or more e-books stored in association with content feature information for which a similarity equal to or less than the threshold has been calculated. Therefore, the suggestion information generator 110 can extract either or both e-books determined to be similar to the user information and e-books determined to be dissimilar to the user information. In the content suggestion system of this embodiment, the suggestion information generator 110 extracts content using a single threshold, but this is not limiting. For example, when using multiple thresholds, various methods can be considered as long as they extract content based on the similarity between the user feature information and the content feature information, such as classifying similarity into multiple levels and extracting only content in a predetermined category.

[0031] The content proposal system 100 can transmit, to the user terminal 200, proposal data display information for displaying the proposal information on the user terminal 200 via the proposal information transmission unit 111 (S307). The proposal data display information is transmitted to the user terminal 200 as an HTML file, and is intended to display the proposal information on the display screen of the user terminal 200. The proposal data display information may also be intended to display the proposal information on an application installed on the user terminal 200. As a result, the user terminal 200 displays proposal information including one or more e-books. The proposal information may also be arranged in order of similarity between a plurality of e-books.

[0032] As described above, the content proposal system 100 can convert user information including the user's emotional information into user characteristic information that can be compared with content characteristic information, using the functions of the content proposal system 100 or an external conversion server 400. The content proposal system 100 can generate proposal information by extracting one or more e-books stored in association with the content characteristic information based on the similarity between the user characteristic information and the content characteristic information, and transmit to the user terminal 200 proposal data display information for displaying the proposal information on the display screen of the user terminal 200. Therefore, when a user performs a content search, the content proposal system 100 can enable one or more e-books that are suited to the user's real-time emotional state to be viewed on the screen of the user terminal 200.

[0033] Furthermore, since the content recommendation system 100 of this embodiment recognizes emotional information from the user's voice, the emotional information is an objective evaluation, and the system can recommend content that is appropriate for emotions that the user himself / herself is unable to subjectively recognize. Therefore, the content recommendation system 100 can support the user in searching for content such as books and movies. [Explanation of symbols]

[0034] 100...content suggestion system, 104...content data storage unit, 105...user information receiving unit, 106...emotion information transmitting unit, 107...emotion information receiving unit, 108...user information transmitting unit, 109...user information receiving unit, 110...suggestion information generating unit, 111...suggestion information transmitting unit, 200...user terminal, 300...analysis server, 400...conversion server

Claims

1. A content suggestion method for making it possible to browse one or more pieces of content suited to a user's emotions from content data that refers to content data in which each piece of content is associated with content feature information generated by converting the data format of each piece of content, the method comprising: receiving user information including at least emotion information of the user from a user terminal of the user; converting the data format of the received user information into user characteristic information that can be compared with the content characteristic information; generating suggested information by referencing the content data and extracting one or more pieces of content stored in association with the content feature information based on a similarity between the user feature information and each piece of content feature information; transmitting, to the user terminal, proposed data display information for displaying the proposed information on the user terminal; A content suggestion method comprising:

2. The content suggestion method according to claim 1 , wherein the user characteristic information and the content characteristic information are numerical vectors.

3. The content suggestion method described in claim 2, characterized in that the suggestion information is generated by determining similarity based on a predetermined threshold and extracting one or more pieces of content stored in association with content feature information that is similar and / or dissimilar to the user feature information.

4. The content suggestion method described in claim 2, characterized in that the conversion to the user characteristic information is performed by sending the user information to a conversion server that converts the user information into a numerical vector, and receiving the user characteristic information generated by the conversion server from the conversion server.

5. 2. The content suggestion method according to claim 1, wherein the emotion information is made up of a plurality of items, each of which has a set value.

6. The content suggestion method according to claim 2 , wherein the user information includes personal attributes that indicate characteristics of the user.

7. The method of claim 2 , wherein the user information includes one or more selection items related to detailed information about the user.

8. 6. The content suggestion method according to claim 1, wherein the user information includes any text data input at the user terminal.

Images