JP7705830B2 - Information processing device, information processing method, and information processing program - Google Patents

Description

The present invention relates to an information processing device, an information processing method, and an information processing program.

Conventionally, there is a service that provides comments posted by users on news articles to other users who have viewed the news articles (see, for example, Patent Document 1).

JP 2019-197422 A

However, the conventional technology leaves room for further improvement in terms of improving the quality of services related to comments.

The present application has been made in consideration of the above, and aims to provide an information processing device, an information processing method, and an information processing program that can improve the quality of services related to comments.

The information processing device according to the present application includes an identification unit and a detection unit. The identification unit identifies a user who agrees with a comment whose content satisfies a predetermined content condition. The detection unit detects other comments whose content satisfies the predetermined condition based on the agreement status of the identified user with other comments.

According to one aspect of the embodiment, it is possible to improve the quality of services related to comments.

FIG. 1A is a diagram showing a first process executed by an information processing apparatus according to an embodiment. FIG. 1B is a diagram showing a second process executed by the information processing apparatus according to the embodiment. FIG. 1C is a diagram showing a third process executed by the information processing apparatus according to the embodiment. FIG. 1D is a diagram showing a fourth process executed by the information processing device according to the embodiment. FIG. 2 is a diagram illustrating an example of the configuration of an information processing system according to the embodiment. FIG. 3 is a diagram illustrating an example of the configuration of the information processing apparatus according to the embodiment. FIG. 4 is a diagram illustrating an example of user information. FIG. 5 is a diagram showing an example of content information. FIG. 6 is a diagram showing an example of comment information. FIG. 7 is a flowchart showing a processing procedure of a first process executed by the information processing apparatus according to the embodiment. FIG. 8 is a flowchart showing the procedure of the second process executed by the information processing apparatus according to the embodiment. FIG. 9 is a flowchart showing the procedure of the third process executed by the information processing apparatus according to the embodiment. FIG. 10 is a flowchart showing the procedure of the fourth process executed by the information processing apparatus according to the embodiment. FIG. 11 is a diagram illustrating an example of a hardware configuration.

Below, the information processing device, information processing method, and information processing program according to the present application will be described in detail with reference to the drawings. Note that the information processing device, information processing method, and information processing program according to the present application are not limited to these embodiments. In addition, the same parts in the following embodiments will be denoted by the same reference numerals, and duplicated descriptions will be omitted.

(Embodiment)
First, the process executed by the information processing device according to the embodiment will be described with reference to Figures 1A to 1D. Figures 1A to 1D are diagrams showing processes 1 to 4 executed by the information processing device according to the embodiment. Figures 1A to 1D show an example of the operation of an information processing system S including an information processing device 1 according to the embodiment.

As shown in Figures 1A to 1D, an information processing system S according to an embodiment includes an information processing device 1, a plurality of user terminals 100, and a requesting terminal 200. Note that, for ease of explanation, the requesting terminal 200 is omitted in Figures 1B to 1D.

In the information processing system S according to the embodiment, various processes are performed based on comments posted by users on specific content, thereby improving the quality of services related to comments.

First, we will explain process 1 shown in FIG. 1A. Specifically, the information processing device 1 first accepts a request for content distribution from a requesting user via the request source terminal 200 (step S1).

The content is, for example, information about a news article in a news distribution service, or information about a product for sale in a shopping service. In other words, the content is content that can accept comments from the user who distributed it. In the case of a news article, the comments are comments on the news article, and in the case of a product, the comments are reviews of the product or reviews of the store that sells the product.

Next, the information processing device 1 delivers the received content to the user via the user terminal 100 (step S2). Specifically, when the information processing device 1 receives a content delivery request from the user terminal 100, it delivers the specified content to the user terminal 100.

Next, the information processing device 1 accepts comments on the content from the user who distributed the content (step S3). The comments are, for example, text format information, but are not limited to this and may be audio format information or image format information. Note that comments may be accepted while the user is in the middle of entering them, or after the user has pressed a button (post button) to request posting after entering them.

The information processing device 1 then analyzes the received comments and identifies the trend of the comment content (step S4). The trend includes, for example, whether the comment content is favorable or unfavorable toward the entire content or a part of the content, or whether the comment content supplements the entire content or a part of the content. The trend may be set in advance, or may be automatically generated based on the comment content.

The trend can be identified, for example, using a classification model in machine learning. Specifically, the information processing device 1 identifies the trend of the comments by inputting the comments into a classification model that has learned the comment contents and trends as a data set.

Next, the information processing device 1 classifies each comment according to the identified tendency (step S5). Specifically, the information processing device 1 sorts each comment according to the identified tendency.

For example, when the number of comments becomes a predetermined number or more, the information processing device 1 stops sorting comments into that trend from the point in time when the number reaches the predetermined number or more. Furthermore, when the information processing device 1 detects a trend in which a predetermined number or more comments are classified in a short period (a predetermined period), it performs a process of thinning out unnecessary comments (e.g., random extraction) from among the predetermined number or more of comments. In other words, when a predetermined number or more of comments are classified into a particular trend, the information processing device 1 considers the predetermined number or more of comments to be fraudulent comments (fake comments, abusive comments) and prevents them from increasing any further.

Next, the information processing device 1 provides various notifications to the user based on the tendency of the comment content (step S6). Specifically, when the identified tendency is similar (including the same) to the tendency of comments that have already been posted a predetermined number or more, the information processing device 1 provides a posting-prohibited notification indicating that the comment cannot be posted. The posting-prohibited notification may be, for example, notified by hiding the posting button displayed on the user terminal 100 when the user is in the middle of inputting a comment, or may be notified by a text message indicating that the posting cannot be accepted after the user has pressed the posting button. That is, the information processing device 1 performs control to avoid accepting the posting of a comment. This makes it possible to avoid, for example, a decrease in the diversity of comments due to a large number of comments with the same tendency being posted. That is, according to the information processing device 1 according to the embodiment, it is possible to increase the diversity of comments, thereby improving the quality of services related to comments.

In addition, the information processing device 1 compares the comment status for each classified trend and suggests posting comments for other trends based on the comparison results. Specifically, the information processing device 1 compares the number of comments for each trend and suggests posting comments for trends where the number of comments is less than a predetermined number. Specifically, if a predetermined number or more comments have already been collected for the posted comment trend, the information processing device 1 suggests posting a comment for a trend where the number of comments is less than the predetermined number. This makes it possible to increase the diversity of comments, thereby improving the quality of services related to comments.

Next, the information processing device 1 determines whether the classification result obtained by the processing of step S5 satisfies a predetermined condition (step S7). Specifically, the information processing device 1 determines whether a predetermined number or more of comments with a certain tendency among a plurality of predefined tendencies have been posted (classified).

Next, when the classification result satisfies a predetermined condition, more specifically, when a predetermined number or more of comments of a plurality of pre-set tendencies have been posted, the information processing device 1 provides the posted comments of each trend together (step S8-1). In other words, rather than providing the posted comments immediately, the comments of various tendencies are provided together at the timing when they are posted. This makes it possible to prevent only comments with a biased tendency from being provided, and since comments of various tendencies can be provided together, it is possible to increase the diversity of comments. In other words, according to the information processing device 1 according to the embodiment, it is possible to improve the quality of services related to comments.

In addition, the information processing device 1 provides the user who posted the comment with comments that have a different trend from the trend of the comment they posted (step S8-2). For example, the information processing device 1 provides a collection of comments with trends excluding the trend of the comment posted by the user from among the comments of each trend. In this way, the posting user can view comments with trends other than the trend of the comment posted by the posting user. In other words, according to the information processing device 1 according to the embodiment, it is possible to improve the quality of services related to comments.

Next, we will explain the process 2 shown in Figure 1B. Note that in Figure 1B, for the sake of convenience, the description of processes that overlap with the processes shown in Figure 1A may be omitted.

In the second process shown in FIG. 1B, the information processing device 1 collects (obtains) evaluation information indicating an evaluation of each comment posted on the content (step S31). The evaluation information includes whether the comment is favorably evaluated (agreeing with the comment) or negatively evaluated (disagreeing with the comment).

Next, the information processing device 1 judges the diversity of evaluations based on the acquired evaluation information (step S32). For example, if the evaluations for the comment are only positive (or only negative), the information processing device 1 judges that there is no diversity of evaluations. In other words, if the number of evaluations (number of users) for both positive and negative evaluations is balanced (if the difference in the number of evaluations is less than a predetermined value), the information processing device 1 judges that there is diversity of evaluations.

Furthermore, the information processing device 1 determines the diversity of the evaluations included in the set for each combination of multiple comments. Specifically, the information processing device 1 determines whether or not there is diversity in the evaluations of the comments included in the set. For example, when user A evaluates comment A and comment B, and gives a favorable evaluation for comment A and a negative evaluation for comment B, the information processing device 1 determines that there is diversity in the evaluations of the set of comment A and comment B. Note that, although the diversity is determined based on the evaluation of user A in the above, the diversity may also be determined based on the evaluation of a group including multiple users with similar user information. Note that the number of comments included in the set may be three or more.

In addition, the information processing device 1 determines that there is diversity in evaluations for a group that includes comments with only positive evaluations and comments with only negative evaluations. In other words, when the evaluations are viewed on a group-by-group basis, the information processing device 1 determines that there is diversity in evaluations if there is a balance between positive evaluations and negative evaluations.

Then, the information processing device 1 determines whether or not to display the comment (whether or not to display the comment) based on the determination result (step S33). For example, the information processing device 1 determines the comment determined to have evaluation diversity as the comment to be displayed. In addition, the information processing device 1 determines the comment of the group determined to have evaluation diversity as the comment to be displayed.

Next, the information processing device 1 displays the comments (set of comments) determined as the display targets on the user terminal 100 (step S34). In this way, according to the information processing device 1 according to the embodiment, it is possible to improve the quality of services related to comments by providing comments with a diversity of evaluations.

Next, we will explain the process 3 shown in Figure 1C. Note that in Figure 1C, for the sake of convenience, the description of processes that overlap with the processes shown in Figures 1A and 1B may be omitted.

In process 3 shown in FIG. 1C, the information processing device 1 collects each comment on the content (step S51).

Then, the information processing device 1 extracts comments whose content satisfies a predetermined content condition (step S52). For example, the information processing device 1 extracts comments whose content is inappropriate (score is a predetermined value or more) using a model trained on feature information extracted from the comment content and a score indicating the degree of inappropriateness of the comment content as a data set.

Next, the information processing device 1 identifies users who agree with the extracted comment (step S53). Specifically, the information processing device 1 identifies users who have given favorable ratings to the extracted comment.

Next, the information processing device 1 detects other comments whose comment content meets a predetermined condition based on the agreement status of the identified user (specific user) with other comments (step S54). For example, the information processing device 1 detects other comments where the number of specific users who agree with the other comments is a predetermined number or more as other comments that meet the predetermined condition. In other words, for other comments whose comment content is determined to be not inappropriate in step S52, the information processing device 1 detects other comments with a large number of specific users as comments whose comment content is inappropriate. This makes it possible to find other inappropriate comments from the agreement status of users who agree with the inappropriate comment, thereby improving the quality of services related to comments.

Next, the information processing device 1 displays the comments on the user terminal 100, excluding the other comments detected in step S54 and the comments extracted in step S52 from the display targets (step S55).

Next, we will explain process 4 shown in Figure 1D. Note that in Figure 1D, for the sake of convenience, the description of processes that overlap with the processes shown in Figures 1A to 1C may be omitted.

In process 4 shown in FIG. 1D, the information processing device 1 distributes content to the user terminal 100 (step S71). Next, the information processing device 1 accepts comments on the content from the user (step S72).

Next, the information processing device 1 classifies each received comment according to the tendency of the comment content or according to the user information of the posting user (step S73).

Next, the information processing device 1 identifies user information corresponding to the comments viewed by the viewing user, or a trend of the comments (step S74). Specifically, the information processing device 1 identifies user information or a trend into which the comments previously viewed by the viewing user are classified.

Next, the information processing device 1 extracts comments that correspond to user information different from the identified user information, or comments that have a different tendency from the identified tendency (step S75).

Then, the information processing device 1 provides the extracted comments to the user terminal 100 of the viewing user (step S76).

This allows us to provide comments with different trends and user information than the comments viewed by the viewing user, improving the quality of the service related to comments.

Next, a configuration example of an information processing system S according to an embodiment will be described with reference to FIG. 2. FIG. 2 is a block diagram showing a configuration example of an information processing system S according to an embodiment. As shown in FIG. 2, in the information processing system S according to an embodiment, an information processing device 1, a plurality of user terminals 100, and a plurality of requesting terminals 200 are connected to a network N by wire or wirelessly. The network N is, for example, the Internet, a Wide Area Network (WAN), a Local Area Network (LAN), or the like.

The information processing device 1 is a server device that executes the information processing method according to the embodiment. The information processing device 1 executes the processes 1 to 4 described above with reference to Figures 1A to 1D.

In addition, the information processing device 1 is an information processing device that cooperates with multiple user terminals 100 and multiple requesting terminals 200, and provides each user terminal 100 and each requesting terminal 200 with API (Application Programming Interface) services for various applications (hereinafter, apps), etc., and various data, and is realized by a server device, a cloud system, etc.

The information processing device 1 may also be an information processing device that provides some kind of web service online to each user terminal 100 and each requesting terminal 200. For example, the information processing device 1 may provide services such as Internet connection, search services, SNS (Social Networking Service), electronic commerce (EC), electronic payment, online games, online banking, online trading, accommodation and ticket reservations, video and music distribution, news, maps, route searches, route guidance, line information, operation information, and weather forecasts as web services. In practice, the information processing device 1 may cooperate with various servers that provide the above-mentioned web services, and may act as an intermediary for the web services or be responsible for processing the web services.

The user terminal 100 is a terminal device owned by a user who views content, comments, and ratings on comments. The user terminal 100 can be any type of terminal device, such as a smartphone, a desktop PC, a notebook PC, or a tablet PC. The user terminal 100 transmits various types of information to the information processing device 1, etc., and receives information provided by the information processing device 1, etc.

The requesting terminal 200 is a terminal device owned by a requesting user who requests the distribution of content. The requesting terminal 200 can be any type of terminal device, such as a smartphone, a desktop PC, a notebook PC, or a tablet PC. The requesting terminal 200 transmits various information to the information processing device 1, etc., and receives information provided by the information processing device 1, etc.

Next, an example configuration of the information processing device 1 will be described with reference to FIG. 3.

Fig. 3 is a diagram showing an example of the configuration of an information processing device 1 according to an embodiment. As shown in Fig. 3, the information processing device 1 has a communication unit 2, a control unit 3, and a memory unit 4. The control unit 3 has a distribution unit 31, an acquisition unit 32, an identification unit 33, a classification unit 34, a determination unit 35, a decision unit 36, a detection unit 37, a reception control unit 38, a suggestion unit 39, and a provision unit 40. The memory unit 4 stores user information 41, content information 42, and comment information 43.

The communication unit 2 is realized, for example, by a network interface card (NIC). The communication unit 2 is connected to the network via a wired or wireless connection.

The control unit 3 is a controller, and is realized, for example, by a processor such as a CPU (Central Processing Unit) or MPU (Micro Processing Unit) that executes various programs (corresponding to an example of an information processing program) stored in a storage device inside the information processing device 1 using a RAM or the like as a working area. The control unit 3 is also a controller, and may be realized, for example, by an integrated circuit such as an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a GPGPU (General Purpose Graphic Processing Unit).

The storage unit 4 is realized, for example, by a semiconductor memory element such as a random access memory (RAM) or a flash memory, or a storage device such as a hard disk or an optical disk.

User information 41 is information about a user. FIG. 4 is a diagram showing an example of user information 41. As shown in FIG. 4, user information 41 includes items such as "user ID" and "user information."

"User ID" is identification information that identifies a user. "User information" is information about a user. "User information" includes, for example, attribute information about the user's attributes and behavioral information. Attribute information includes psychographic attributes and demographic attributes. Behavioral information includes the user's behavior on the network (search behavior, purchasing behavior, comment posting behavior, rating posting behavior, etc.).

Next, content information 42 is information related to content. FIG. 5 is a diagram showing an example of content information 42. As shown in FIG. 5, content information 42 includes items such as "content ID," "type," "content information," and "comment information."

"Content ID" is identification information that identifies the content. "Type" is information that indicates the type of content. "Content information" is information that indicates the content, and includes text information, image information, etc. "Comment information" is information on comments posted on the content, and the "Comment ID" of comment information 43 described later is input.

Next, comment information 43 is information about comments posted on content. FIG. 6 is a diagram showing an example of comment information 43. As shown in FIG. 6, comment information 43 includes items such as "comment ID," "tendency," "content information," and "rating information."

"Comment ID" is identification information that identifies a comment. "Trend" is information that indicates the trend of the comment content. "Content information" is information that indicates the comment content, and includes text information. "Evaluation information" is information about the evaluations made by users of the comment, and includes, for example, information about the number of users who gave favorable evaluations and the number of users who gave unfavorable evaluations.

Next, the functions of the control unit 3 of the information processing device 1 (the distribution unit 31, the acquisition unit 32, the identification unit 33, the classification unit 34, the judgment unit 35, the decision unit 36, the detection unit 37, the reception control unit 38, the proposal unit 39, and the provision unit 40) will be described.

The distribution unit 31 distributes content to the user terminal 100. The distribution unit 31 acquires the content to be distributed from the requesting user via the requesting terminal 200, and stores the content in the content information 42. Then, when the distribution unit 31 receives a content distribution request from the user via the user terminal 100, it distributes the target content to the user terminal 100.

In addition to the content, the distribution unit 31 also distributes comments posted by other users who viewed the content, and ratings posted by other users on the comments.

The acquisition unit 32 acquires comments on the content from the user who distributed the content. The timing of comment acquisition by the acquisition unit 32 is arbitrary. For example, the acquisition unit 32 acquires the comment being input when a predetermined condition is satisfied during comment input (the number of characters is a predetermined number or more, the elapsed time since input is a predetermined time or more, etc.). The acquisition unit 32 also acquires the comment when the above-mentioned post button is pressed after the comment is input. The acquisition unit 32 also acquires evaluation information indicating the evaluation of the comment.

The identification unit 33 identifies a trend in the comment content based on the comment content. The trend can be identified, for example, using a classification model in machine learning. Specifically, the identification unit 33 identifies the trend in the comment by inputting the comment into a classification model that has learned the comment content and trends as a data set.

The identification unit 33 also identifies user information of the user who posted the comment. The identified user information is, for example, attribute information such as gender, location, and occupation, and behavioral information such as search behavior and purchasing behavior.

The identification unit 33 also identifies users who agree with a comment whose content is determined by the determination unit 35 (described later) to satisfy a predetermined content condition. For example, the identification unit 33 identifies users who have provided favorable evaluations of comments that satisfy a predetermined content condition.

The identification unit 33 also identifies a tendency corresponding to a posting user based on the behavioral history of the posting user who posted the comment. The behavioral history of the comment is, for example, a history of past comment postings and a history of evaluations of comments posted by other posting users.

For example, the identification unit 33 identifies a trend corresponding to the posting user based on the trends of comments posted in the past by the posting user. Specifically, the identification unit 33 counts the trends of comments posted in the past, and identifies the trend with the largest number of comments as the trend corresponding to the posting user. Furthermore, the identification unit 33 may identify multiple trends with a predetermined number or more of comments as trends (multiple) corresponding to the posting user based on the above counting results.

The identification unit 33 also identifies a tendency corresponding to the posting user based on the contents of the evaluation history for comments posted by other posting users. Specifically, when a posting user gives a favorable evaluation to a comment posted by another posting user, the identification unit 33 identifies the tendency of the evaluated comment as a tendency corresponding to the posting user.

In addition, when a posting user gives a negative evaluation to a comment posted by another posting user, the identification unit 33 identifies a tendency different from the tendency of the evaluated comment as a tendency corresponding to the posting user.

In addition, as a method for selecting a trend different from the trend of the rated comment, it is possible to select based on the distance between trends in the distributed representation space. In other words, the identification unit 33 identifies a trend that is a predetermined distance or more away from the trend of the rated comment in the distributed representation space as the trend corresponding to the posting user.

The classification unit 34 classifies each comment according to the identified tendency or according to the identified user information. Specifically, the classification unit 34 sorts each comment according to the identified tendency or according to the identified user information. More specifically, the classification unit 34 classifies the identified tendency of the comment into one of a number of trends set in advance. Such classification is possible by inputting the identified tendency and the tendency to which the identified tendency belongs among the multiple trends set in advance into a model that has been trained as a data set.

For example, for a trend or user information in which the number of comments exceeds a predetermined number, the classification unit 34 stops sorting comments into that trend or user information once the predetermined number has been reached. In addition, for trends or user information into which a predetermined number or more comments have been classified in a short period (predetermined period), the classification unit 34 performs a process of thinning out unnecessary comments (e.g., random extraction) from among the predetermined number or more of comments. In other words, when a predetermined number or more of comments have been classified into a particular trend or user information, the classification unit 34 considers the predetermined number or more of comments to be fraudulent comments (fake comments, abusive comments) and prevents them from increasing any further.

The classification unit 34 may also re-set the multiple trends to be classified based on the classification results. Specifically, when the number of comments classified into a specific trend is significantly large, in other words, when the number of comments classified into a specific trend is equal to or greater than a predetermined number and the number of comments classified into other trends is less than a predetermined number, the classification unit 34 may analyze the comment content of the classified comments and subdivide the specific trend into multiple trends. Alternatively, the classification unit 34 may reduce the number of other trends by combining other trends with fewer than a predetermined number of comments.

Such a method of reconstructing trends can be achieved, for example, by creating a distributed representation of each comment based on the comment content, and automatically generating axes that divide areas in the distributed representation space as trends.

The determination unit 35 performs various determination processes. For example, the determination unit 35 determines whether or not the trend of the identified comment satisfies a condition based on the trend of already posted comments. Specifically, the determination unit 35 determines that the above condition is satisfied when the trend of the identified comment is similar to the trend into which a predetermined number of posted comments are classified.

The determination unit 35 also calculates the ratio of the number of posted comments for each trend to the number of posted comments for all trends, and determines that the above condition is met if the trend for which this ratio is equal to or greater than a predetermined value is similar to the identified comment trend.

The determination unit 35 may also determine whether the above conditions are met by taking into account the relationship between the posting user and the tendencies of comments posted by the posting user in the past. Specifically, the determination unit 35 first estimates a risk indicating the likelihood of posting for each trend, depending on the number of comments for each trend posted in the past by the posting user. The risk is set higher for a trend with a larger number of comments. The determination unit 35 then calculates a score indicating the degree of similarity between the identified comment tendency and the tendencies into which a predetermined number of posted comments are classified, and determines whether the above conditions are met based on the score and the above risk.

The determination unit 35 also determines the diversity of evaluations based on the evaluation information acquired by the acquisition unit 32. For example, the determination unit 35 determines that there is no diversity of evaluations when the evaluations for the comment are only favorable evaluations (or only negative evaluations). In other words, the determination unit 35 determines that there is diversity of evaluations when the number of evaluations (number of users) for both favorable evaluations and negative evaluations is balanced (when the difference in the number of evaluations is less than a predetermined value).

The determination unit 35 may also determine the diversity of evaluations in a plurality of comments. For example, the determination unit 35 determines the diversity of evaluations included in a set for each combination of a plurality of comments. Specifically, the determination unit 35 determines whether or not there is diversity in the evaluations of the comments included in the set. For example, when user A evaluates comment A and comment B, if the user A evaluates comment A favorably and evaluates comment B negatively, the determination unit 35 determines that there is diversity in the evaluations of the set of comment A and comment B. Note that, although the diversity is determined based on the evaluation of user A in the above, the diversity may also be determined based on the evaluation of a group including a plurality of users with similar user information. Note that the number of comments included in a set may be three or more.

The determination unit 35 also determines that there is diversity in evaluations for a group that includes comments with only positive evaluations and comments with only negative evaluations. In other words, when the evaluations are viewed on a group-by-group basis, the determination unit 35 determines that there is diversity in evaluations when there is a balance between positive evaluations and negative evaluations.

The determination unit 35 may also determine the diversity of users who have made evaluations. For example, the determination unit 35 determines that there is diversity in evaluations when the attribute information of each user who has made a favorable evaluation (or a negative evaluation) of a comment is not similar, that is, when users with various attributes have made favorable evaluations.

On the other hand, when the attribute information of each user who gave a favorable (or negative) rating to a comment is similar, i.e., when users with similar attributes give favorable ratings, the determination unit 35 determines that there is no diversity in the ratings.

The determination unit 35 may also group users based on user information and determine the diversity of evaluations on a group-by-group basis. For example, the determination unit 35 determines that there is diversity of evaluations when the information (e.g., group attribute information) of groups that have given favorable (or negative) evaluations to a comment is not similar, that is, when groups with various attributes have given favorable evaluations.

On the other hand, when the information of each group that gave a favorable (or negative) evaluation to a comment is similar, i.e., when groups with similar attributes give favorable evaluations, the judgment unit 35 judges that there is no diversity in the evaluations.

The determination unit 35 also determines whether the comment content satisfies a predetermined content condition. Specifically, the determination unit 35 inputs the comment into a model that has been trained using feature information extracted from the comment content and a score indicating the degree of inappropriateness of the comment content as a data set, and determines that the content condition is satisfied if the score output from the model is equal to or greater than a predetermined value. In other words, the content condition is a condition for determining whether the comment content is inappropriate (violent expression, obscene expression, defamatory expression, etc.). The feature information also includes the presence or absence of violent, obscene, or defamatory expression sentences contained in the comment content, words related to these expressions, and the number of such sentences and words. The determination unit 35 may correct the above model based on other comments that satisfy the predetermined condition (inappropriate) detected by the detection unit 37 described later.

The determination unit 36 determines whether or not to display a comment based on the determination result of the determination unit 35. Specifically, the determination unit 36 determines whether or not to display a comment based on the determination result of the diversity of evaluations. Specifically, the determination unit 36 determines a comment determined to have diversity of evaluations as a display target. In addition, the determination unit 36 determines a group of comments determined to have diversity of evaluations as a display target.

The determination unit 36 also determines the display order of the comments based on the judgment result of the diversity of evaluations. For example, the determination unit 36 displays comments that are judged to have diversity of evaluations at the top of a plurality of comments that have a large number of users who have given favorable (or negative) evaluations to the comments.

The detection unit 37 detects other comments whose comment content satisfies a predetermined condition based on the agreement status of the users (users who agree with the comment that satisfies the content condition) identified by the identification unit 33 to the other comments. Specifically, the detection unit 37 detects other comments whose number of users who agree with the other comment among the identified users is a predetermined number or more as other comments that satisfy the predetermined condition. Alternatively, the detection unit 37 may detect other comments whose ratio of the number of users identified by the identification unit 33 to the number of all users who agree with the other comment is a predetermined value or more as other comments that satisfy the predetermined condition. In other words, the detection unit 37 detects other comments whose comment content is determined by the determination unit 35 to be not inappropriate and whose majority of the identified users agree as comments whose comment content is inappropriate. Note that the other comments may be comments posted for the same content (e.g., the same news article) as the comment that satisfies the predetermined content condition, or may be comments posted for other content of the same type (e.g., other news articles). Furthermore, the detection unit 37 may detect appropriate comments from among the comments determined to be inappropriate based on the agreement state of the identified user with other comments. Furthermore, the other comments detected by the detection unit 37 may be comments seeking sympathy from other users, comments posted by the parties involved in the content, etc., in addition to inappropriate comments.

The detection unit 37 may also detect other comments that satisfy a predetermined condition by taking into account the context information of the identified user. For example, when the context information of the user when agreeing with an inappropriate comment is similar to the context information of the user when agreeing with another comment, the detection unit 37 detects the other comment as another comment that satisfies a predetermined condition. That is, the detection unit 37 estimates from the user's context information whether the user feels like agreeing with the inappropriate comment (feeling contrarian), and detects other comments that the user agrees with when he or she feels like agreeing with the inappropriate comment as other (inappropriate) comments that satisfy a predetermined condition. Note that the context information includes a context related to a search such as a search query, and a context related to the user's emotions. The context related to the user's emotions can be inferred from, for example, the written expression of an email, the conversational expression of a voice conversation, etc.

The reception control unit 38 performs control to avoid accepting comments based on the judgment result of the judgment unit 35. Specifically, when the identified tendency is similar (including the same) as the tendency of comments that have already been posted a predetermined number or more, the reception control unit 38 issues a posting-prohibited notification indicating that the comment cannot be posted. The posting-prohibited notification may be issued, for example, by hiding the post button displayed on the user terminal 100 when the user is in the middle of inputting a comment, or may be issued as a text message indicating that the post cannot be accepted after the user has pressed the post button.

The reception control unit 38 may also suggest changing the trend of the input comments as a control to avoid accepting comments. In such a case, the reception control unit 38 suggests changing the trend to one in which the number of classified comments is less than a predetermined number, based on the classification results of the classification unit 34.

The suggestion unit 39 compares the comment status for each trend classified by the classification unit 34, and suggests posting a comment based on the comparison result. The suggestion to post a comment may be made, for example, when viewing content, or when entering a comment.

For example, the suggestion unit 39 compares the number of comments for each trend and suggests posting comments for which the number of comments tends to be less than a predetermined number. The suggestion unit 39 also compares the ratio of the number of comments for each trend to the total number of comments, which is the sum of the number of comments for each trend, and suggests posting comments for which the ratio tends to be less than a predetermined value. Alternatively, when the difference in the ratio between the trends is equal to or greater than a predetermined value, the suggestion unit 39 suggests posting comments for which the ratio tends to be low.

Furthermore, when the number of trends is less than a predetermined number, the suggestion unit 39 suggests posting a comment with a new trend. Furthermore, when suggesting posting a comment with a specific trend, the suggestion unit 39 suggests auxiliary information that supports posting of the comment with the specific trend based on previously posted comments with the specific trend. The auxiliary information is, for example, summary data of the posted comments, or list data in which words related to a specific trend are extracted from the posted comments and listed.

Furthermore, when the number of comments for each trend is balanced, and there exists a trend in which the number of comments whose content meets a predetermined content condition is equal to or greater than a predetermined number, the suggestion unit 39 suggests posting a comment for that trend. In other words, when a certain number of comments for each trend has been secured, but there are many comments with inappropriate content for a particular trend, the suggestion unit 39 suggests posting a comment for that particular trend. Specifically, when a predetermined number or more comments have already been collected for the posted comment trend, the suggestion unit 39 suggests posting a comment for a trend in which the number of comments is less than the predetermined number.

The providing unit 40 provides various information to the user. For example, the providing unit 40 provides comments of each trend when the classification result by the classification unit 34 satisfies a predetermined condition. Specifically, the providing unit 40 provides the posted comments of each trend in a lump, assuming that the predetermined condition is satisfied when a predetermined number or more of the multiple trends set in advance are classified into. After providing the comments in a lump, the providing unit 40 resets the classification result, reclassifies newly posted comments, and provides the newly posted comments of each trend in a lump when the above-mentioned predetermined condition is satisfied. In other words, the providing unit 40 provides the comments of each trend in a lump each time the classification result satisfies the predetermined condition.

The comments provided may be all comments for each trend, or for a trend in which there is a predetermined number or more of comments, one or more representative comments may be selected from the predetermined number or more of comments. The providing unit 40 may select based on the posting history of the posting user, for example, by giving priority to comments from a posting user who has posted a large number of comments in the past, or may simply select at random.

The providing unit 40 may also select a representative comment based on the similarity of attribute information between the posting user and the viewing user. Specifically, the providing unit 40 selects a comment posted by a posting user that has a high similarity to the attribute information of the viewing user as the representative comment.

The providing unit 40 also provides a viewing user who has viewed a comment with one of the tendencies classified by the classification unit 34 with a comment with a different tendency from the trend. The providing unit 40 also provides a viewing user who has viewed a comment corresponding to one of the user information classified by the classification unit 34 with a comment corresponding to the user information different from the user information.

For example, the providing unit 40 creates a distributed representation of each trend (or each piece of user information) and provides a comment of the trend (or user information) that is the furthest from the trend (or user information) of the comment viewed by the viewing user in the distributed representation space. The providing unit 40 may also provide comments of multiple trends (or user information) that are a predetermined distance or more away from the trend (or user information) of the comment viewed by the viewing user. The providing unit 40 may also select a representative comment by the above method when the number of comments of the trend to be provided is a predetermined number or more.

In addition, the providing unit 40 provides the posting user who posted the comment with a comment that has a different tendency from the tendency of the posted comment. In other words, the providing unit 40 provides the posting user with a comment that has a different tendency from the tendency corresponding to the posting user identified by the identifying unit 33 based on the behavioral history of the posting user.

For example, the providing unit 40 hides comments with a tendency corresponding to the posting user and displays comments with a tendency different from the tendency. In addition, the providing unit 40 arranges the display order of comments with a tendency different from the tendency corresponding to the posting user higher than the comment with the tendency corresponding to the posting user.

In addition, the providing unit 40 displays comments with a tendency corresponding to the posting user and comments with a tendency different from the tendency in different display modes. The display modes include, for example, background color, character color, character underline, character font, character size, etc.

The providing unit 40 also provides information about other comments that satisfy the predetermined conditions (that are inappropriate) detected by the detecting unit 37 to, for example, an administrator who manages the content (or comments).

Next, the processing procedures of processes 1 to 4 executed by the information processing device 1 according to the embodiment will be described with reference to Figs. 7 to 10. Figs. 7 to 10 are flowcharts showing the processing procedures of processes 1 to 4 executed by the information processing device 1 according to the embodiment.

First, the process 1 will be described with reference to FIG. 7. As shown in FIG. 7, the control unit 3 first distributes content to the user terminal 100 (step S101).

Next, the control unit 3 accepts user comments on the distributed content (step S102).

Next, the control unit 3 identifies trends in the comment content (step S103).

Next, the control unit 3 classifies the comments according to the identified trends (step S104).

Next, the control unit 3 determines whether the identified tendency is similar to the tendency of posted comments that have been posted a predetermined number of times or more (step S105).

If the identified tendency is not similar to the tendency of posted comments that have been posted a predetermined number of times or more (step S105: No), the control unit 3 determines whether the classification result satisfies a predetermined condition (step S106).

If the classification results satisfy the predetermined condition (step S106: Yes), the control unit 3 compiles the comments for each trend and provides them to the user (step S107), and ends the process.

On the other hand, if the classification result does not satisfy the predetermined condition (step S106: No), the control unit 3 ends the process without providing a comment.

In addition, in step S105, if the identified tendency is similar to the tendency of posted comments that have been posted a predetermined number of times or more (step S105: Yes), the control unit 3 notifies the user that the comment cannot be posted or suggests posting a comment of a different tendency (step S108), and ends the process.

Next, the second process will be described with reference to FIG. 8. As shown in FIG. 8, the control unit 3 collects evaluation information indicating evaluations of the comments (step S201).

Next, the control unit 3 determines the diversity of the evaluations based on the collected evaluation information (step S202).

Next, the control unit 3 determines the comments to be displayed based on the judgment result (step S203).

The control unit 3 then provides the determined comment to the user (step S204) and ends the process.

Next, the third process will be described with reference to FIG. 9. As shown in FIG. 9, the control unit 3 collects each user's comment on the content (step S301).

Next, the control unit 3 extracts comments whose content satisfies predetermined content conditions (step S302).

The control unit 3 then identifies users who agree with the extracted comment (step S303).

Then, the control unit 3 detects other comments based on the identified user's agreement with other comments (step S304) and ends the process.

Next, the fourth process will be described with reference to FIG. 10. As shown in FIG. 10, the control unit 3 classifies each comment according to the tendency of the comment content (or according to user information) (step S401).

Next, the control unit 3 identifies the trends (user information) of the comments viewed by the viewing user (step S402).

The control unit 3 then extracts comments that have a different tendency (or user information) from the identified tendency (or user information) (step S403).

The control unit 3 then provides the extracted comments to the user (step S404) and ends the process.

〔others〕
In addition, among the processes described in the above embodiments, some of the processes described as being performed automatically can be performed manually. Alternatively, all or some of the processes described as being performed manually can be performed automatically using a known method. In addition, the information including the processing procedures, specific names, various data, and parameters shown in the above documents and drawings can be changed arbitrarily unless otherwise specified. For example, the various information shown in each drawing is not limited to the illustrated information.

In addition, each component of each device shown in the figure is a functional concept, and does not necessarily have to be physically configured as shown in the figure. In other words, the specific form of distribution and integration of each device is not limited to that shown in the figure, and all or part of them can be functionally or physically distributed and integrated in any unit depending on various loads, usage conditions, etc.

For example, a part or all of the memory unit 4 shown in FIG. 3 may be stored in a storage server or the like, rather than being stored in each device. In this case, each device obtains various pieces of information by accessing the storage server.

[Hardware configuration]
The information processing device 1 according to the embodiment described above is realized by a computer 1000 having a configuration as shown in Fig. 11, for example. Fig. 11 is a diagram showing an example of a hardware configuration. The computer 1000 is connected to an output device 1010 and an input device 1020, and has a configuration in which a calculation device 1030, a primary storage device 1040, a secondary storage device 1050, an output IF (Interface) 1060, an input IF 1070, and a network IF 1080 are connected by a bus 1090.

The arithmetic device 1030 operates based on programs stored in the primary storage device 1040 and the secondary storage device 1050, programs read from the input device 1020, and the like, and executes various processes. The primary storage device 1040 is a memory device, such as a RAM, that temporarily stores data used by the arithmetic device 1030 for various calculations. The secondary storage device 1050 is a storage device in which data used by the arithmetic device 1030 for various calculations and various databases are registered, and is realized by a ROM (Read Only Memory), a HDD (Hard Disk Drive), a flash memory, or the like.

The output IF 1060 is an interface for transmitting information to be output to an output device 1010 that outputs various types of information, such as a monitor or printer, and is realized, for example, by a connector conforming to a standard such as USB (Universal Serial Bus), DVI (Digital Visual Interface), or HDMI (registered trademark) (High Definition Multimedia Interface). The input IF 1070 is an interface for receiving information from various input devices 1020, such as a mouse, keyboard, scanner, etc., and is realized, for example, by a USB.

The input device 1020 may be a device that reads information from, for example, an optical recording medium such as a CD (Compact Disc), a DVD (Digital Versatile Disc), or a PD (Phase change rewritable Disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. The input device 1020 may also be an external storage medium such as a USB memory.

The network IF 1080 receives data from other devices via the network N and sends it to the computing device 1030, and also transmits data generated by the computing device 1030 to other devices via the network N.

The arithmetic unit 1030 controls the output device 1010 and the input device 1020 via the output IF 1060 and the input IF 1070. For example, the arithmetic unit 1030 loads a program from the input device 1020 or the secondary storage device 1050 onto the primary storage device 1040 and executes the loaded program.

For example, when the computer 1000 functions as the information processing device 1, the arithmetic unit 1030 of the computer 1000 realizes the functions of the control unit 3 by executing a program loaded onto the primary storage device 1040.

〔effect〕
As described above, the information processing device 1 according to the embodiment includes the classification unit 34 and the provision unit 40. The classification unit 34 classifies each comment posted on a predetermined content according to the tendency of the comment content. The provision unit 40 provides the comments of each tendency when the classification result satisfies a predetermined condition.

The information processing device 1 according to the embodiment also includes an identification unit 33, a determination unit 35, and a reception control unit 38. The identification unit 33 identifies a trend in the content of comments posted on a specific piece of content that are entered by users. The determination unit 35 determines whether the trend of the identified comment satisfies a condition based on the trend of comments that have already been posted. The reception control unit 38 performs control to avoid the reception of comment postings based on the determination result.

The information processing device 1 according to the embodiment also includes an acquisition unit 32, a determination unit 35, and a decision unit 36. The acquisition unit 32 acquires evaluation information indicating evaluations of comments posted on a specific piece of content. The determination unit 35 determines the diversity of the evaluations based on the acquired evaluation information. The decision unit 36 determines whether or not to display the comments based on the determination result.

The information processing device 1 according to the embodiment also includes a classification unit 34 and a provision unit 40. The classification unit 34 classifies each comment posted on a specific piece of content according to the tendency of the comment content. The provision unit 40 provides a viewing user who has viewed a comment with one of the classified tendencies with a comment with a different tendency from the comment.

The information processing device 1 according to the embodiment also includes a classification unit 34 and a provision unit 40. The classification unit 34 classifies each comment posted by a posting user on a specific piece of content according to the posting user's user information. The provision unit 40 provides a viewing user who has viewed a comment corresponding to any of the classified user information with a comment corresponding to user information different from the user information.

The information processing device 1 according to the embodiment also includes a classification unit 34, a specification unit 33, and a provision unit 40. The classification unit 34 classifies each comment posted on a specific piece of content according to the tendency of the comment content. The specification unit 33 specifies a tendency corresponding to the user from among the classified tendencies based on the user's behavior history regarding comments. The provision unit 40 provides the user with comments with a tendency different from the specified tendency.

The information processing device 1 according to the embodiment also includes an identification unit 33 and a detection unit 37. The identification unit 33 identifies a user who agrees with a comment whose content satisfies a predetermined content condition. The detection unit 37 detects other comments whose content satisfies the predetermined condition based on the agreement status of the identified user with other comments.

The information processing device 1 according to the embodiment also includes a classification unit 34 and a suggestion unit 39. The classification unit 34 classifies each comment posted on a specific piece of content according to the tendency of the comment content. The suggestion unit 39 compares the comment status for each classified tendency and suggests posting a comment based on the comparison result.

The information processing device 1 according to each of the above-described embodiments can improve the quality of services related to comments.

Although several embodiments of the present application have been described in detail above with reference to the drawings, these are merely examples, and the present invention can be embodied in other forms that incorporate various modifications and improvements based on the knowledge of those skilled in the art, including the forms described in the disclosure section of the invention.

〔others〕
In addition, among the processes described in the above embodiments, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically by a known method. In addition, the information including the processing procedures, specific names, various data and parameters shown in the above documents and drawings can be changed arbitrarily unless otherwise specified. For example, the various information shown in each drawing is not limited to the illustrated information.

In addition, each component of each device shown in the figure is a functional concept, and does not necessarily have to be physically configured as shown in the figure. In other words, the specific form of distribution and integration of each device is not limited to that shown in the figure, and all or part of them can be functionally or physically distributed and integrated in any unit depending on various loads, usage conditions, etc.

In addition, the processes described in the above embodiments can be combined as appropriate to the extent that they do not cause any contradictions in the process content.

The above-mentioned "section, module, unit" can be read as "means" or "circuit." For example, the control unit 3 can be read as control means or control circuit.

Reference Signs List 1 Information processing device 2 Communication unit 3 Control unit 4 Storage unit 31 Distribution unit 32 Acquisition unit 33 Identification unit 34 Classification unit 35 Determination unit 36 Determination unit 37 Detection unit 38 Acceptance control unit 39 Proposal unit 40 Provision unit 41 User information 42 Content information 43 Comment information 100 User terminal 200 Requester terminal S Information processing system

Claims (5)

an acquisition unit that acquires comments on a predetermined content from a user who has distributed the content;
A determination unit that determines whether the comment content is inappropriate by using a model that has been trained using feature information extracted from the comment content of the comment and a score indicating the degree of inappropriateness of the comment content as a data set; and
An identification unit that identifies a specific user who has given a favorable evaluation to the comment;
a detection unit that detects , from among the other comments whose content is determined to be not inappropriate, the other comments whose number of the specific users is equal to or greater than a predetermined number as the inappropriate comments ,
The determination unit is
The model is corrected using learning data in which a correct label of an inappropriate comment is assigned to the other comment that is determined by the model to be not the inappropriate comment and that is detected as the inappropriate comment by the detection unit.
Information processing device. The detection unit is
The information processing device according to claim 1 , wherein the other comment is detected as the inappropriate comment when a ratio of the number of users identified by the identification unit to the number of all users who agree with the other comment is equal to or greater than a predetermined value. The detection unit is
The information processing device according to claim 1 , wherein if context information of the user when agreeing with the inappropriate comment is similar to context information of the user when agreeing with another comment, the other comment is detected as the inappropriate comment. 1. A computer-implemented information processing method, comprising:
an acquisition step of acquiring comments on a predetermined content from a user who has distributed the content;
a determination step of determining whether the comment content is inappropriate, using a model trained on a data set of feature information extracted from the comment content of the comment and a score indicating the degree of inappropriateness of the comment content;
A step of identifying a specific user who has given a favorable evaluation to the comment ;
and a detection step of detecting , from among the other comments whose content is determined to be not inappropriate, the other comments whose number of specific users is equal to or exceeds a predetermined number as the inappropriate comments ,
The determination step includes:
The model is corrected using learning data in which a correct label of an inappropriate comment is assigned to the other comment that is determined by the model to be not the inappropriate comment and that is detected as the inappropriate comment by the detection step.
Information processing methods. An acquisition step of acquiring comments on a given content from a user who has distributed the content;
a judgment step of judging an inappropriate comment, which is a comment whose content is inappropriate, by using a model trained on a data set of feature information extracted from the content of the comment and a score indicating the degree of inappropriateness of the comment content;
A step of identifying a specific user who has given a favorable rating to the comment;
a detection step of detecting , from among the other comments whose content is determined to be not inappropriate, the other comments whose number of the specific users is equal to or exceeds a predetermined number as the inappropriate comments ;
The determination procedure includes:
The model is corrected using learning data in which a correct label of an inappropriate comment is assigned to the other comment that is determined by the model to not be the inappropriate comment and that is detected as the inappropriate comment by the detection step.
Information processing program.

Images