JP7061328B1 - Information processing equipment, information processing systems and programs - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information processing device, an information processing system and a program for determining the future influence of posting on SNS. In an information processing system (1), an information processing apparatus (2) uses a target person information acquisition unit (21) for acquiring target person information for identifying a target person, and the target person information. A keyword collection unit (22) that collects keywords related to the target person, a keyword search processing unit (23) that searches for posts containing keywords from posts on the SNS using keywords, and search results for posts containing keywords. The post extraction unit (25) that extracts the target person-related posts that are related to the target person, and the target person-related posts whose future diffusion status is predicted to satisfy the attention condition from the target person-related posts. It is provided with a diffusion prediction unit (26) for extracting attention posts, and a notification unit (27) for notifying an external terminal of attention posts. [Selection diagram] Fig. 1

Description

The present invention relates to an information processing apparatus, an information processing system and a program.

In recent years, in social networking services (hereinafter, simply referred to as "SNS"), the number of posts that are increasing day by day is enormous due to the increase in the number of users worldwide. Some of these posts have attracted a lot of people's attention over time and, in some cases, have become more socially recognized. For example, when a post that mentions a certain group or individual attracts the attention of many people, the impact of the post on society or the like can have a great impact on the activities of the group or individual. Therefore, it has become important for various groups or individuals to monitor posts on SNS by performing so-called "egosearch" and the like.

Here, for example, in Patent Document 1, the disclosure unit of an individual's opinion is extracted from the content information collected from the Web, and the disclosure unit of the extracted individual's opinion is analyzed to be the subject of the individual's opinion. A content information processing method that identifies an individual's evaluation of is disclosed. According to this content information processing method, further, by using the reference degree for each disclosure unit of the content information, it is possible to search for a noteworthy opinion on the Web having a high degree of influence.

Japanese Patent No. 4097602

In the conventional technique described in Patent Document 1, the degree of influence of the content information up to the present can be determined by using the reference degree. However, with such a conventional technique, there is a problem that it is not possible to determine the future impact of posting on SNS.

It is an object of the present invention to solve the above problems and to obtain an information processing apparatus, an information processing system and a program capable of determining the future influence of posting on the SNS.

The information processing apparatus according to the present invention includes a target person information acquisition unit that acquires target person information for identifying a target person, a keyword collection unit that collects keywords related to the target person using the target person information, and a keyword collection unit. A post that extracts target person-related posts that are related to the target person by using the keyword search processing unit that searches for posts that include the keyword from posts on the SNS using keywords and the search results of posts that include the keyword. An extraction unit, a diffusion prediction unit that extracts attention posts that are target person-related posts whose future diffusion status is predicted to satisfy the attention condition from the target person-related posts, and a notification unit that notifies the attention post to an external terminal. , Equipped with.

According to the present invention, a post containing a keyword is searched from posts on SNS using a keyword related to the target person collected using the target person information for identifying the target person, and a search result of the post including the keyword is searched. Is used to extract posts related to the target audience. Since the target person-related posts whose future diffusion status is expected to satisfy the attention condition are extracted from the target person-related posts as the attention posts, the information processing apparatus according to the present invention has a future influence on the posts on the SNS. You can judge.

It is a block diagram which shows the structure of the information processing system which concerns on Embodiment 1. FIG. It is a block diagram which shows the hardware configuration which realizes the function of the information processing apparatus which concerns on Embodiment 1. FIG. It is a flowchart which shows the operation of the information processing apparatus which concerns on Embodiment 1. FIG. It is a figure which shows an example of the post with an image attached. It is a graph which shows the time change of the number of retweets of a post in SNS. It is a screen diagram which shows an example of an alert notification screen.

Embodiment 1.
FIG. 1 is a block diagram showing a configuration of an information processing system 1 according to the first embodiment. In FIG. 1, the information processing system 1 is a system in which an information processing device 2, an SNS server 3, a Web server 4, a service user terminal 5, and an SNS user terminal 6 are connected via a network 7. The information processing system 1 is used to predict the future impact of posting on SNS. The prediction of future impact is specifically a prediction of the future spread of posts (hereinafter also referred to as “spread prediction”). The network 7 is a telecommunication line including the Internet. The information processing device 2 is a server that provides the service user terminal 5 with a diffusion prediction service that predicts the future diffusion status of posts corresponding to the target person from posts on the SNS.

The SNS server 3 is a server that provides an SNS to the SNS user terminal 6 via the network 7. Examples of SNS include Twitter (registered trademark), Instagram (registered trademark), TikTok (registered trademark), YouTube (registered trademark), and Facebook (registered trademark). The web server 4 is a server that provides a website via the network 7. A large number of Web servers 4 are connected to the network 7, and the Web server 4 discloses a huge amount of information. The service user terminal 5 and the SNS user terminal 6 are terminal devices capable of communicating with the information processing device 2, the SNS server 3 or the Web server 4 via the network 7, and are, for example, a smartphone, a tablet terminal, or a PC (Personal Computer). Is.

The service user terminal 5 is a terminal used by a user who uses the diffusion prediction service provided by the information processing apparatus 2. Users of the spread prediction service can be various groups or individuals. "Groups" include various commercial or non-profit organizations such as corporations, sole proprietorships or administrative organizations, and "individuals" include various natural persons such as politicians, entertainers, and the general public. Is also included. Further, the user of the diffusion prediction service may be the "target person" itself, which will be described later, or may be other than the "target person". In the service user terminal 5, a diffusion prediction service provided in the form of SaaS (Software as a Service) can be used. That is, the service application for providing the spread prediction service is executed on the information processing apparatus 2, and the service user terminal 5 is the spread prediction service on the Web browser without installing the service-dedicated application. You can receive the offer. Alternatively, a dedicated application for using the diffusion prediction service may be installed in the service user terminal 5. The SNS user terminal 6 is a terminal used by a user who uses SNS (hereinafter referred to as "SNS user"). For example, the SNS user can use the SNS by logging in to the SNS provided by the SNS server 3 using the SNS user terminal 6.

The information processing device 2 that provides the diffusion prediction service includes a target person information acquisition unit 21, a keyword collection unit 22, a keyword search processing unit 23, a design search processing unit 24, a post extraction unit 25, a diffusion prediction unit 26, and a notification unit 27. Be prepared. The target person information acquisition unit 21 acquires the target person information for identifying the target person. The "target person" can be various groups or various individuals. "Groups" include various commercial or non-profit organizations such as corporations, sole proprietorships or administrative organizations, and "individuals" include various natural persons such as politicians, entertainers, and the general public. Is also included. For example, a user of the diffusion prediction service (hereinafter, also referred to as a “service user”) inputs target person information to the service user terminal 5 on which the input screen provided by the service application is displayed. The target person information acquisition unit 21 acquires the target person information from the service user terminal 5 via the network 7.
If the target person is a group, the target person information is, for example, the current name that can identify the group. For example, if the target person is a non-profit organization, the target person information is the current organization name of the non-profit organization, and if the target person is a company, the target person information is the current company name of the company. possible. If the subject is an individual, the subject information is, for example, the current name that can identify the individual. For example, if the target person is a general person, the target person information may be the current name of the person, and if the target person is a celebrity, the target person information may be the current stage name of the celebrity.

The keyword collection unit 22 collects keywords related to the target person by using the target person information. For example, when the target person is a company, the keyword collection unit 22 searches the Web page of the website published by the Web server 4 via the network 7 by using the target person information such as the current company name. , Collect keywords in search results. The website for collecting keywords includes, for example, an Internet encyclopedia site such as Wikipedia (registered trademark) or a website provided by the target person. For example, when the target person is a company and the target for collecting keywords is the website of the company, the keywords are collected from a page such as an IR (Investor Relations) page where comprehensive information of the company can be acquired.

Further, as keywords, the keyword collection unit 22 includes, in addition to the current name of the target person, the product name of the product related to the target person, the service name of the service related to the target person, and the product name of the product related to the target person. Abbreviation, abbreviation of service name of service related to target person, category of product related to target person, category of service related to target person, past name of target person, current abbreviation of target person, past of target person Collect information that includes at least one of the abbreviations and information that indicates the industry to which the subject belongs.
Products related to the target person include, for example, products provided by the company (hereinafter referred to as "target company") for a fee or free of charge when the target person is a company, and the target person is an individual. In this case, an article traded by the individual (hereinafter referred to as "target individual") using a so-called flea market application or the like, a literary work created by the target individual, or a work in which the target individual appears is included.
Services related to the target person include, for example, services provided by the target company for a fee or free of charge, or volunteer activities performed by the target individual.
The product name includes, for example, the name of each product as well as the brand name of the product, and the service name includes, for example, the name of each service as well as the brand name of the service.
The name of the target person includes, for example, the company name of the target company, the name of the target individual, or another name of the individual.
The industry to which the target person belongs includes, for example, the industry to which the target company belongs or the occupation of the target individual.

The keywords collected by the keyword collecting unit 22 are registered in a database stored in a storage device (not shown) included in the information processing device 2. When an operation requesting the display of a keyword is input to the service user terminal 5 in the information processing system 1, the information processing device 2 acquires the registered keyword from the database and transmits it to the service user terminal 5. .. When an editing operation for editing the keyword is input after the keyword is displayed on the screen of the service user terminal 5, the information processing apparatus 2 uses the editing content instructed by the editing operation for the keyword registered in the database. And edit. Normally, the keyword collecting unit 22 collects a plurality of keywords. The editing contents that can be instructed by the editing operation are addition of a new keyword to these plurality of keywords, deletion of one of the keywords, correction of any of the keywords, and the like.

The keyword search processing unit 23 searches for posts containing the keyword from the posts on the SNS using the keyword. For example, the keyword search processing unit 23 performs at least one of an OR search, an AND search, a phrase search, or an excluded word search for each keyword. In the OR search, it is possible to search for posts containing any of a plurality of keywords related to the target person. In the AND search, it is possible to search for posts that include all of a plurality of keywords related to the target person. In addition, the keyword search not only searches for the text information included in the SNS post, but also recognizes the text included in the attached image and attached video included in the SNS post, and searches for the recognized text. You may search.

In the phrase search, keywords related to the target person can be searched as one phrase. In the excluded word search, unnecessary words can be excluded from the search target while keeping the keywords related to the target person originally desired to be searched. For example, when a post containing the keyword "cafe" is the search target, "net cafe" having a different meaning from "cafe" can be excluded from the search target.

In addition, the keyword search processing unit 23 can also search for posts in which the keywords are written in hidden letters. For example, the keyword search processing unit 23 searches for a post in which the keyword is described in the hidden character by analyzing the language of the post and estimating the keyword from the hidden character. In order to prevent the keyword "A news agency" from being searched, if the post is described in the hidden letters "A news agency", the keyword search processing unit 23 analyzes the post in a language and starts from "A news agency". By estimating "A news agency", posts that do not directly describe "A news agency" are searched.

The design search processing unit 24 searches for a post including a design related to the target person by using an image included in the post on the SNS. Some posts on SNS do not include keywords related to the target person in the text, but have an image containing a design related to the target person attached. Posts with the image attached may also be spread on SNS to the extent that it affects the target audience. When the information processing apparatus 2 includes the design search processing unit 24, such posts can be searched. For example, the design search processing unit 24 analyzes the image attached to the post and searches for the post to which the image including the design is attached. The image may be a still image, a moving image, or both. In addition, as the design related to the target person, for example, a logo such as a symbol mark indicating the target person, a trademark including a design character or a figure used by the target person, a brand of a product or a service provided by the target person is shown. The logo, the shape of the product provided by the target person, the shape of the package of the product provided by the target person, the logo attached to the product or package provided by the target person, the UI (User Interface) of the application provided by the target person, etc. be.

Further, the design search processing unit 24 can also search for posts including the design from the posts including the keyword indicated by the search result by the keyword search processing unit 23 among the posts in the SNS. For example, the design search processing unit 24 targets posts that include keywords related to the target person in the text, which are searched by the keyword search processing unit 23, and the design search processing unit 24 attaches an image containing a design related to the target person. search for.

The design search processing unit 24 searches for a post including a design by using a design-learned model that outputs a result indicating the presence or absence of a design by inputting an image. The design trained model is generated by machine learning that applies the learning algorithm of image recognition. The learning data used for the machine learning is preferably a coarse image that can be attached to an actual post on the SNS. It is possible to generate a design-learned model with improved design inference accuracy by using an image whose size is equal to or larger than a certain value as training data. Further, in the learning of the design-learned model, data augmentation may be performed on the image which is the existing learning data to increase the number of training data. In the data augmentation, new training data can be generated by performing image inversion, image angle conversion, image blurring processing, etc., which are existing training data.

The post extraction unit 25 extracts the target person-related post, which is a post related to the target person, by using the search result of the post including the keyword by the keyword search processing unit 23.
For example, the post extraction unit 25 extracts all the posts including the keyword searched by the keyword search processing unit 23 as the target person-related posts as they are.
Further, for example, the post extraction unit 25 classifies the search results of posts including the keyword into categories, and extracts the target person-related posts belonging to the category set by the service user. Categories include, for example, campaign posts, press releases, service quality opinions, employee response opinions, product reviews, product or service requests, product introductions, service introductions, breakdowns, foreign matter contamination, information leaks, service failure complaints, etc. There are operational mistakes, store congestion, traffic congestion, reputational damage, comparisons with other companies, or information related to infectious diseases.

Further, the post extraction unit 25 may extract the target person-related posts by using the search result of the post including the keyword by the keyword search processing unit 23 and the search result of the post including the design by the design search processing unit 24. ..
For example, the post extraction unit 25 attaches an image including a design related to the target person by the design search processing unit 24 to the post containing a specific keyword among the posts including the keyword searched by the keyword search processing unit 23. Only the posts searched as the posted posts are extracted as the target-related posts. At that time, posts containing keywords other than specific keywords are extracted as they are as target-related posts. In such an extraction method, first, a search using a broader keyword as a part of multiple keywords is performed to search the post widely so that there is no omission in the search, and the broader keyword is used. It is effective as a method to extract only the posts related to the target person by searching only the posts with the image containing the design related to the target person and excluding the irrelevant posts.
Further, for example, the post extraction unit 25 classifies the search result of the post including the keyword and the search result of the post including the design into each category, and extracts the target person-related posts belonging to the category set by the service user. This makes it possible to extract posts containing designs related to the target person as posts related to the target person in addition to posts containing the keyword.

The diffusion prediction unit 26 extracts the attention posts, which are the target-related posts whose future diffusion status is predicted to satisfy the attention condition, from the target-related posts. The attention condition is a condition indicating that the subject-related posts are widely spread. For example, when the degree of change in the value indicating the actual diffusion status in the predetermined observation period after the subject-related post is posted, the diffusion prediction unit 26 exceeds the predetermined degree of change, the subject. We predict that the future spread of related posts will meet the above-mentioned attention conditions. The value indicating the diffusion status is, for example, at least one of the number of reposts, the number of citations, the number of evaluations, and the number of views of the subject-related posts. Further, the diffusion prediction unit 26 calculates a predicted arrival value of a value indicating a future diffusion situation, and when the calculated predicted arrival value exceeds a predetermined threshold value, the future diffusion situation satisfies the attention condition. May be predicted.

A notable contributor is a contributor whose posts can be directly viewed by a considerable number of SNS users. For example, if the SNS is Twitter, the contributor who is followed by a considerable number of SNS users is the contributor of interest. Such contributors are also called influencers. The future spread of posts by featured contributors is expected to meet the conditions of interest, and at least the posts by featured contributors are directly viewed by a significant number of SNS users. Therefore, the diffusion prediction unit 26 predicts that the future diffusion situation satisfies the attention condition when the subject-related post is reposted, quoted, or evaluated by the attention contributor.

The diffusion prediction unit 26 also uses a diffusion-learned model that inputs information related to the target person and outputs information indicating whether or not the future diffusion status satisfies the attention condition, and the future diffusion status determines the attention condition. It may be one that predicts whether or not it is satisfied. For example, the diffusion-learned model is a machine learning model that inputs feature information extracted from a subject-related post and outputs an inference result as to whether or not the future diffusion situation satisfies the attention condition. Various information can be used as the characteristic information of the subject-related posts. As characteristic information, for example, a profile such as a poster's post name, the number of SNS users who follow a person who spreads a poster or a target-related post by quoting, etc., and a so-called "bot" that the poster automatically posts by a program. Information indicating whether or not, a specific word in the target person-related post, information indicating whether or not the target person-related post has positive or negative content, an image attached to the target person-related post, re-currently One or a combination of the number of posts, the number of citations, the number of evaluations, the number of views, etc., the elapsed time from the posting of the target person-related post to the present time, or the number of reposts per unit time of the target person-related post. Can be used.
The diffusion prediction unit 26 may also calculate the transition of the predicted value indicating the future diffusion status of the attention post, in addition to extracting the attention post.

The notification unit 27 notifies the service user terminal 5, which is an external terminal, of the attention posting. When the diffusion prediction unit 26 calculates the transition of the predicted value, the notification unit 27 provides control information for displaying the attention posting and the transition of the predicted value on the screen of the service user terminal 5. The service user terminal 5 may be notified by transmitting to. In this case, the service user terminal 5 displays the attention posting and the transition of the predicted value on the screen according to the control information received from the information processing apparatus 2. It should be noted that a part of the transition of the predicted value in this case may include the actual transition of the value indicating the diffusion status up to the present time. Further, the form of the notification may be, for example, a notification on the provision screen of the diffusion prediction service, a notification by pop-up display, a notification by transmission by e-mail, or a notification by voice.

FIG. 2 is a block diagram showing a hardware configuration that realizes the functions of the information processing apparatus 2. For example, the information processing apparatus 2 has a communication interface 100, an input / output interface 101, a processor 102, and a memory 103 as hardware configurations. Each function of the target person information acquisition unit 21, the keyword collection unit 22, the keyword search processing unit 23, the design search processing unit 24, the post extraction unit 25, the diffusion prediction unit 26, and the notification unit 27 provided in the information processing apparatus 2 is provided. It is realized by these hardware configurations.

The communication interface 100 outputs the data received from the external device via the network 7 to the processor 102, and transmits the data generated by the processor 102 to the service user terminal 5 via the network 7. The processor 102 reads and writes data to and from a storage device (not shown) in FIGS. 1 and 2 via the input / output interface 101. The storage device is a storage device included in a computer that functions as an information processing device 2.

The application program for realizing each function of the target person information acquisition unit 21, the keyword collection unit 22, the keyword search processing unit 23, the design search processing unit 24, the post extraction unit 25, the diffusion prediction unit 26, and the notification unit 27 is It is stored in the storage device included in the computer that functions as the information processing device 2. The processor 102 reads the application program stored in the storage device, loads it into the memory 103, and executes the loaded program via the input / output interface 101. As a result, the processor 102 realizes the functions of the target person information acquisition unit 21, the keyword collection unit 22, the keyword search processing unit 23, the design search processing unit 24, the post extraction unit 25, the diffusion prediction unit 26, and the notification unit 27. do. The memory 103 is, for example, a RAM (Random Access Memory).

FIG. 3 is a flowchart showing the operation of the information processing apparatus 2.
The target person information acquisition unit 21 acquires the target person information for identifying the target person (step ST1). The keyword collecting unit 22 collects keywords related to the target person by using the target person information (step ST2). The keywords collected by the keyword collecting unit 22 are registered in the database in the storage device (not shown) included in the information processing device 2 as they are or after being edited by the service user.

The keyword search processing unit 23 uses the keywords read from the database to search for posts containing the keywords from the posts on the SNS (step ST3). FIG. 4 is a diagram showing an example of a post 10 to which an image 10B is attached. The keyword search processing unit 23 searches for the post 10 in which the keyword is included in the body 10A of the post 10. Further, the design search processing unit 24 searches for the post 10 in which the design 10C relating to the target person is included in the image 10B. These processes are the search processes in step ST3.

The post extraction unit 25 extracts the target person-related posts using the search results of the posts by the search process (step ST4). The diffusion prediction unit 26 extracts the attention posts, which are the target-related posts whose future diffusion status is predicted to satisfy the attention condition, from the target-related posts (step ST5). FIG. 5 is a graph showing the time change of the number of retweets (hereinafter referred to as “RT number”) of the tweets posted in the SNS when the SNS is Twitter. In FIG. 5, the value indicating the future diffusion status of the post is the RT number of the post.

When the degree of change in the actual number of RTs in the predetermined observation period ΔT1 after the subject-related post extracted by the post extraction unit 25 is posted, the diffusion prediction unit 26 exceeds a predetermined threshold value. It is predicted that the future diffusion status of the target person-related post will satisfy the attention condition, and the target person-related post will be extracted as the attention post. In the noteworthy post shown in FIG. 5, the number of RTs increases logarithmically by the time time ΔT2 (> ΔT1) elapses after posting. The diffusion prediction unit 26 determines whether or not the degree of change in the actual number of RTs of the post during the observation period ΔT1 is larger than the threshold value, and thus whether or not the number of RTs of the post exceeds a certain number after ΔT2 has elapsed. Can be predicted. In general, the time for ΔT2 until the increase in the number of RTs slows down and approaches the maximum value, that is, the increase reaches a plateau, is several days, weeks, or months depending on the content of the post. There are various units.

The change in the number of RTs of the subject-related posts can be predicted, for example, based on the degree of change in the actual number of RTs in the observation period ΔT1 after the subject-related posts are posted. For example, the diffusion prediction unit 26 calculates a predicted arrival value at which the increase in the number of RTs of the target-related post reaches a plateau, and when the predicted arrival value exceeds the threshold value, the future diffusion status of the post satisfies the attention condition. Predict. It should be noted that a plurality of threshold values may be prepared in advance, the importance of the attention posting may be determined according to which threshold value the predicted arrival value exceeds, and the notification may be made via the notification unit 27.

The notification unit 27 notifies the service user terminal 5 of the attention post extracted by the diffusion prediction unit 26 (step ST6). For example, the notification unit 27 transmits control information for displaying an alert notification screen including information about a noteworthy post to the service user terminal 5 via the network 7. The service user terminal 5 displays an alert notification screen regarding the attention posting according to the control information received from the notification unit 27.

FIG. 6 is a screen view showing an example of the alert notification screen 30, and shows a case where the post shown in FIG. 4 is a noteworthy post. As shown in FIG. 6, the alert notification screen 30 displays the attention posting 10, the RT arrival predicted value 30A, and the transition graph 30B of the number of RTs. The service user can know the degree of influence of the attention post on the target person by referring to the attention post 10, the RT arrival predicted value 30A, and the transition graph 30B of the number of RTs on the alert notification screen 30.

Further, the operations of step ST1 and step ST2 in FIG. 3 are usually performed only once when a new target person is registered. However, the keyword editing work can be performed at any time. On the other hand, the operations from step ST3 to step ST6 in FIG. 3 are repeatedly executed in a predetermined cycle. That is, the keyword search processing unit 23, the post extraction unit 25, the diffusion prediction unit 26, and the notification unit 27 repeatedly perform processing at a predetermined cycle. The cycle of the iterative process is, for example, a one-minute cycle.

As described above, the information processing apparatus 2 according to the first embodiment uses the keywords related to the target person collected by using the target person information for identifying the target person, and posts including the keywords from the posts on the SNS. Search for, and use the search results of posts that include keywords to extract target person-related posts, and from target person-related posts, target person-related posts that are expected to meet the attention conditions in the future diffusion status are set as attention posts. Since it is extracted, the information processing apparatus 2 can determine the future influence of the posting on the SNS.

In the information processing apparatus 2 according to the first embodiment, the keyword collecting unit 22 collects keywords from a website. Keywords can be collected from the huge amount of information published as a website.

In the information processing apparatus 2 according to the first embodiment, the keyword collecting unit 22 uses the current name of the target person, the product name of the product related to the target person, and the service name of the service related to the target person as keywords. , Abbreviation of product name of product related to target person, abbreviation of service name of service related to target person, category of product related to target person, category of service related to target person, past name of target person, Collect information that includes at least one of the subject's current abbreviations, the subject's past abbreviations, and information indicating the industry to which the subject belongs. As a result, the information processing apparatus 2 can reliably extract posts related to the target person from the SNS.

The information processing apparatus 2 according to the first embodiment includes a design search processing unit 24 that searches for a post including a design related to the target person by using an image included in the post in the SNS. The post extraction unit 25 extracts the target person-related post by using the search result of the post including the keyword and the search result of the post including the design. By including the design search processing unit 24, the information processing apparatus 2 can extract posts related to the target person by using posts including keywords and posts including designs.

In the information processing apparatus 2 according to the first embodiment, the design search processing unit 24 searches for posts including the design from the posts including the keyword indicated by the search result by the keyword search processing unit 23 among the posts in the SNS. As a result, the information processing apparatus 2 can search widely for posts containing keywords from the huge amount of information published as a website, and then narrow down the search to posts containing designs related to the target person. can.

In the information processing apparatus 2 according to the first embodiment, the design search processing unit 24 searches for posts including a design by using a design-learned model that outputs a result indicating the presence or absence of a design by inputting an image. By using a design-learned model, it is possible to accurately search for posts that include designs.

In the information processing apparatus 2 according to the first embodiment, the diffusion prediction unit 26 predicts that the future diffusion situation satisfies the attention condition when the subject-related post is reposted, quoted, or evaluated by the attention contributor. .. As a result, the diffusion prediction unit 26 can accurately extract the target-related posts whose future diffusion status satisfies the attention condition as attention posts.

In the information processing apparatus 2 according to the first embodiment, the diffusion prediction unit 26 predeterminedly determines the degree of change in the value indicating the actual diffusion status in the predetermined observation period after the subject-related post is posted. If the degree of change is exceeded, it is predicted that the future diffusion situation will meet the conditions of interest. As a result, the diffusion prediction unit 26 can accurately extract the target-related posts whose future diffusion status satisfies the attention condition as attention posts.

In the information processing apparatus 2 according to the first embodiment, the diffusion prediction unit 26 calculates a predicted arrival value of a value indicating a future diffusion situation, and when the calculated predicted arrival value exceeds a predetermined threshold value, It is predicted that the future diffusion situation will meet the conditions of interest. As a result, the diffusion prediction unit 26 can accurately extract the target-related posts whose future diffusion status satisfies the attention condition as attention posts.

In the information processing apparatus 2 according to the first embodiment, the diffusion prediction unit 26 uses at least one of the number of reposts, the number of citations, the number of evaluations, and the number of views of the subject-related posts as values indicating the diffusion status. As a result, the diffusion prediction unit 26 can accurately extract the target-related posts whose future diffusion status satisfies the attention condition as attention posts.

In the information processing apparatus 2 according to the first embodiment, the diffusion prediction unit 26 uses a diffusion-learned model that inputs information related to the target person and outputs information indicating whether or not the future diffusion status satisfies the attention condition. Therefore, it is predicted whether or not the future diffusion situation satisfies the attention condition. As a result, the diffusion prediction unit 26 can accurately extract the target-related posts whose future diffusion status satisfies the attention condition as attention posts.

In the information processing apparatus 2 according to the first embodiment, the diffusion prediction unit 26 calculates the transition of the predicted value. The notification unit 27 notifies the service user terminal 5 by transmitting control information for displaying the attention posted and the transition of the predicted value on the screen of the service user terminal 5 to the service user terminal 5. As a result, the service user can grasp the future diffusion status of the featured post.

In the information processing apparatus 2 according to the first embodiment, the keyword search processing unit 23, the post extraction unit 25, the diffusion prediction unit 26, and the notification unit 27 repeatedly perform processing at a predetermined cycle. As a result, the diffusion prediction unit 26 can quickly extract the attention posts whose future diffusion status satisfies the attention condition.

In the information processing system 1 according to the first embodiment, when an operation requesting the display of a keyword is input to the service user terminal 5, the information processing apparatus 2 acquires the keyword registered from the database and the service user terminal. Send to 5. After the keyword is displayed on the screen of the service user terminal 5, when an editing operation for editing the keyword is input, the information processing apparatus 2 uses the editing content instructed by the editing operation for the keyword registered in the database. And edit. As a result, the information processing apparatus 2 can edit the keywords to be more useful.

It is possible to modify any component of the embodiment or omit any component of the embodiment.

1 Information processing system, 2 Information processing device, 3 Service user terminal, 4 SNS user terminal, 4 Web server, 5 SNS server, 10 posts, 10A, 30A posts, 10B image, 10C design, 21 Target person information acquisition unit, 22 Keyword collection unit, 23 Keyword search processing unit, 24 Design search processing unit, 25 Post extraction unit, 26 Diffusion prediction unit, 27 Notification unit, 30 Alert notification screen, 31ART arrival prediction number, 31B transition graph, 100 communication interface, 101 input / output interface, 102 processor, 103 memory.

Claims (16)

Target person information acquisition department that acquires target person information to identify the target person,
A keyword collection unit that collects keywords related to the target person using the target person information,
A keyword search processing unit that searches for posts containing the keyword from posts on social networking services using the keyword.
A post extraction unit that extracts target person-related posts that are related to the target person using the search results of posts containing the keyword, and a post extraction unit.
From the target person-related posts, the diffusion prediction unit that extracts the attention posts that are the target person-related posts whose future diffusion status is predicted to satisfy the attention condition, and
An information processing device including a notification unit that notifies an external terminal of the attention-grabbing post. The information processing apparatus according to claim 1, wherein the keyword collecting unit collects the keywords from a website. In addition to the current name of the target person, the keyword collecting unit may use the product name of the product related to the target person, the service name of the service related to the target person, and the product related to the target person as the keyword. Product name abbreviation, service name abbreviation of service related to the target person, product category related to the target person, service category related to the target person, past name of the target person, said target person The information processing according to claim 1 or 2, wherein the information is collected including at least one of the current abbreviation of the subject, the past abbreviation of the subject, and the information indicating the industry to which the subject belongs. Device. It is provided with a design search processing unit for searching a post containing a design related to the target person by using an image included in the post in the social networking service.
Any of claims 1 to 3, wherein the post extraction unit extracts the target person-related post by using the search result of the post including the keyword and the search result of the post including the design. The information processing apparatus according to item 1. The fourth aspect of the present invention is characterized in that the design search processing unit searches for posts including the design from the posts including the keyword indicated by the search result by the keyword search processing unit among the posts in the social networking service. The information processing device described. Claim 4 or claim 5 is characterized in that the design search processing unit searches for posts including the design by using a design-learned model that inputs the image and outputs a result indicating the presence or absence of the design. The information processing device described in. From claim 1, the diffusion prediction unit predicts that the future diffusion situation satisfies the attention condition when the subject-related post is reposted, quoted, or evaluated by the attention contributor. The information processing apparatus according to any one of claims 6. The diffusion prediction unit is described in the future when the degree of change in the value indicating the actual diffusion status in the predetermined observation period after the posting related to the subject exceeds the predetermined degree of change. The information processing apparatus according to any one of claims 1 to 7, wherein the diffusion state is predicted to satisfy the above-mentioned attention condition. The diffusion prediction unit calculates a predicted arrival value of a value indicating the future diffusion situation, and when the calculated predicted arrival value exceeds a predetermined threshold value, the future diffusion situation determines the attention condition. The information processing apparatus according to any one of claims 1 to 7, wherein the information processing apparatus is predicted to be satisfied. Claim 8 or claim 9 is characterized in that the diffusion prediction unit uses at least one of the number of reposts, the number of citations, the number of evaluations, and the number of views of the subject-related post as a value indicating the diffusion status. The information processing device described in. The diffusion prediction unit uses a diffusion-learned model that inputs information related to the target person and outputs information indicating whether or not the future diffusion status satisfies the attention condition, and the future diffusion status is described. The information processing apparatus according to any one of claims 1 to 7, wherein it predicts whether or not the condition of interest is satisfied. The diffusion prediction unit calculates a transition of a predicted value of a value indicating the diffusion status.
The notification unit is characterized in that it notifies the external terminal by transmitting control information for displaying the attention post and the transition of the predicted value on the screen of the external terminal to the external terminal. The information processing apparatus according to any one of claims 1 to 11. Any one of claims 1 to 12, wherein the keyword search processing unit, the post extraction unit, the diffusion prediction unit, and the notification unit repeatedly perform processing at a predetermined cycle. The information processing apparatus according to item 1. The information processing apparatus according to any one of claims 1 to 13.
The database in which the above keywords are registered and
With the external terminal
When the target person information is input to the external terminal, the target person information acquisition unit acquires the target person information, and the keyword collection unit acquires and collects the keyword using the target person information. Register the said keyword in the database,
An information processing system characterized by this. When an operation requesting the display of the keyword is input to the external terminal,
The information processing device acquires the keyword registered from the database and transmits it to the external terminal.
After the keyword is displayed on the screen of the external terminal, when an editing operation for editing the keyword is input,
The information processing system according to claim 14, wherein the information processing apparatus edits the keyword registered in the database by using the editing content instructed by the editing operation. Computer,
Target person information acquisition department, which acquires target person information to identify the target person,
A keyword collection unit that collects keywords related to the target person using the target person information,
A keyword search processing unit that searches for posts containing the keyword from posts on social networking services using the keyword.
A post extraction unit that extracts target person-related posts that are related to the target person using the search results of posts containing the keyword.
Diffusion prediction unit, which extracts attention posts that are the target person-related posts whose future diffusion status is predicted to satisfy the attention condition from the target person-related posts.
Notification unit that notifies the external terminal of the above-mentioned attention post,
A program to function as.

Images