JP7528713B2 - Apparatus and program for analyzing collected content - Google Patents

Description

The present invention relates to a collected content analysis device and program.

With the widespread use of the Internet, various digital contents (hereinafter referred to simply as "content") such as electronic comics (digital manga), animations, movies, and other videos can now be viewed by downloading or streaming from websites, etc.
On the other hand, there are sites that upload illegal content, such as so-called pirate sites. When illegal content is uploaded to pirate sites, publishers, distributors, copyright holders, etc. of the content are unable to obtain profits (such as copyright fees) that they would otherwise receive, which has become a social problem.

In view of this situation, for example, a system described in Patent Document 1 is disclosed. Patent Document 1 describes an illegal content search system including: a plurality of websites storing web information; and a crawling server that accesses the websites to collect first illegal web information including at least one syllable corresponding to an original keyword among syllables of a first variant keyword included in the web information, divides the first variant keyword of the first illegal web information into phonemes or into phonemes and special characters to generate a second variant keyword in which phonemes excluding special characters are combined in order, determines whether the second variant keyword matches the original keyword, and if they match, classifies the first illegal web information including the second variant keyword matching the original keyword as second illegal web information, and the crawling server accesses the websites using at least one of a plurality of unique authority information having access authority to the websites.

Special table 2019-526109 publication

There is a need to search for illegal content, such as by the method described in US Pat. No. 6,399,633. It is desirable to find such illegal content sooner after it is uploaded to a website.
On the other hand, those who upload illegal content are likely uploading popular content in order to have it viewed by a larger audience, so it seems likely that there is a relationship between illegal content and the demand for that content.

The present invention aims to provide a collected content analysis device and program capable of analyzing illegal content and providing valuable information.

The present invention solves the above problems by the following solving means.
A first invention is a collected content analysis device comprising: an additional collection means for collecting additional content added to a website from a website where content can be viewed and/or downloaded; a legitimate content discrimination means for determining whether the additional content collected by the additional collection means is legitimate content by referring to a legitimate content information storage unit that stores information capable of identifying the legitimate content to be monitored; and an occurrence prediction means for identifying the illegal content to be monitored that is included in the additional content determined to be not legitimate content by the legitimate content discrimination means, analyzing the occurrence status of the illegal content, and predicting trends regarding the occurrence of the illegal content.
A second invention is a collected content analysis device according to the first invention, further comprising a content classification means for classifying the illegal content contained in the additional content that has been determined to be non-legitimate content by the legitimate content discrimination means, and the occurrence prediction means for predicting a tendency for the occurrence of the illegal content for each classification by the content classification means.
A third invention is the collected content analysis device according to the first or second invention, wherein the additional collection means collects the additional content by crawling a specific website.
A fourth invention is a collected content analysis device according to any one of the first to third inventions, further comprising a first evaluation collection means for collecting first posting data about the legitimate content stored in the legitimate content information storage unit from a SNS site, and the occurrence prediction means further uses the first posting data collected by the first evaluation collection means to predict a tendency regarding the occurrence of the illegal content.
A fifth invention is a collected content analysis device, which is any of the collected content analysis devices from the first to fourth inventions, and which comprises a monitoring means for monitoring the website based on a prediction of the illegal content by the occurrence prediction means, and a deletion request means for requesting the deletion of the illegal content from the website when the monitoring means detects the occurrence of the illegal content.
A sixth aspect of the present invention is the collected content analysis apparatus according to the fifth aspect of the present invention, further comprising a request history output unit that outputs deletion request history information made by the deletion request unit.
A seventh invention is a collected content analysis device, which is any of the collected content analysis devices from the first to sixth inventions, and which comprises: a content classification means for classifying the illegal content contained in the additional content that has been determined to be non-legitimate content by the legitimate content discrimination means; and a demand prediction means for predicting demand for content for each classification based on the occurrence status of the illegal content for each classification by the content classification means and the number of views and/or downloads.
An eighth invention is the collected content analysis device of the seventh invention, wherein the content classification means classifies the content by language, and the demand prediction means predicts demand for the content by language.
A ninth invention is a collected content analysis device of the seventh or eighth invention, further comprising a second evaluation collection means for collecting second posting data for each piece of content from a SNS site, and the demand prediction means further uses the second posting data collected by the second evaluation collection means to predict demand for the content.
A tenth invention is a collected content analysis device according to any one of the seventh to ninth inventions, further comprising a demand prediction output means for outputting prediction information of demand for the content by the demand prediction means.
An eleventh aspect of the present invention is a program for causing a computer to function as any one of the collected content analysis devices according to the first to tenth aspects of the present invention.

The present invention provides a collected content analysis device and program that can analyze illegal content and provide valuable information.

1 is an overall configuration diagram of a collected content analysis system according to an embodiment of the present invention and a functional block diagram of a collected content analysis server. 10 is a flowchart showing a content collection process in a collected content analysis server according to the embodiment. 10 is a flowchart showing a posted data collection process in a collected content analysis server according to the embodiment. 10 is a flowchart showing a content monitoring process in a collected content analysis server according to the embodiment; 11 is a diagram showing an example of an occurrence prediction process in a collected content analysis server according to the embodiment; FIG. 13 is a flowchart showing a content demand prediction process in a collected content analysis server according to the embodiment.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that this is merely an example, and the technical scope of the present invention is not limited to this example.
(Embodiment)
<Collected Content Analysis System 100>
FIG. 1 is an overall configuration diagram of a collected content analysis system 100 according to this embodiment and a functional block diagram of a collected content analysis server 1. As shown in FIG.
The collected content analysis system 100 is a system that, in response to a request from a customer, quickly finds illegal content present on a website and provides the customer with a demand forecast for content.

The collected content analysis system 100 includes a collected content analysis server 1 (collected content analysis device), a customer terminal 4, a web server 7, and a SNS (Social Networking Service) server 8. The collected content analysis server 1, the customer terminal 4, the web server 7, and the SNS server 8 are each connected to each other so as to be able to communicate with each other via a communication network N.

<Customer terminal 4>
The customer terminal 4 is, for example, a terminal provided at a publisher, a distribution company, or the like. In the following description, a company such as a publisher or a distribution company that requests content analysis by the collected content analysis system 100 is referred to as a customer. For example, the customer uses the customer terminal 4 to transmit information about content specified by the customer in advance to the collected content analysis server 1, and requests the collected content analysis server 1 to monitor illegal content, etc. The customer terminal 4 also receives information such as content demand forecasts from the collected content analysis server 1.
The customer terminal 4 is, for example, a personal computer (PC), a smartphone, etc. In Fig. 1, the customer terminal 4 is illustrated as a notebook PC.
Although not shown, the customer terminal 4 includes a control unit, a storage unit, an input unit, a display unit, and a communication interface unit. The customer terminal 4 may include a touch panel display in which the input unit and the display unit are integrated.

<Web Server 7>
The web server 7 is a server that provides websites to users, for example, websites that, among all websites around the world, in particular, allow users to view and/or download content.
Although not shown, the web server 7 includes a control unit, a storage unit, and a communication interface unit.

<SNS Server 8>
The SNS server 8 is a server that provides users with an SNS site, which contains, for example, word-of-mouth information, user comments, impressions, and the like.
Although not shown, the SNS server 8 includes a control unit, a storage unit, and a communication interface unit.

<Collected Content Analysis Server 1>
The collected content analysis server 1 is a server operated by, for example, a consulting company, which analyzes content on websites from various angles and provides information on the analysis results.
The collected content analysis server 1 may be configured, for example, by one server, or may be configured by a plurality of servers, or may be a cloud.

As shown in FIG. 2, the collected content analysis server 1 includes a control unit 10 , a storage unit 20 , and a communication interface unit 29 .
The control unit 10 is a CPU (Central Processing Unit) that controls the entire collected content analysis server 1. The control unit 10 appropriately reads and executes an OS (Operating System) and application programs stored in the storage unit 20, thereby cooperating with the above-mentioned hardware and executing various functions.

The control unit 10 includes an additional collection unit 11 (additional collection means), a discrimination unit 12 (legitimate content discrimination means), an evaluation collection unit 13 (first evaluation collection means, second evaluation collection means), an occurrence prediction unit 14 (occurrence prediction means), a monitoring processing unit 15 (monitoring means, deletion request means), a classification unit 16 (content classification means), a demand prediction unit 17 (demand prediction means), and a report output unit 18 (request performance output means, demand prediction output means).

The additional collection unit 11 collects additional content that has been added to a website provided by the web server 7 .
Here, the website from which the additional content collection unit 11 collects additional content is a website where content can be viewed and/or downloaded, as described above with respect to the web server 7.
In addition, the collected content analysis system 100 may limit the websites from which the additional collection unit 11 collects additional content to specific websites, such as “pirated sites” or video-only sites such as YouTube (registered trademark).

The determination unit 12 determines whether the additional content collected by the additional collection unit 11 is legitimate content by referring to the legitimate content information storage unit 22 described below. Here, legitimate content managed by the customer is called legitimate content, while content not managed by the customer and found on a pirated site or the like is called illegal content.

The evaluation collection unit 13 collects, from the SNS site provided by the SNS server 8 , posted data (first posted data) about legitimate content whose information is registered in the legitimate content information storage unit 22 .
Furthermore, the evaluation collection unit 13 collects posted data (second posted data) for each piece of content from the SNS site provided by the SNS server 8 .
Then, the evaluation collection unit 13 stores the collected post data in the post data collection information storage unit 24 described later.

The occurrence prediction unit 14 identifies illegal content to be monitored from the additional content determined by the discrimination unit 12 to be non-legitimate content, analyzes the occurrence status of the identified illegal content, and predicts trends regarding the occurrence of illegal content.
Here, the occurrence prediction unit 14 may specify, for example, content whose title name matches or is similar to that of the legitimate content, among the additional contents determined to be non-legitimate content, as illegal content to be monitored. The occurrence prediction unit 14 may also specify, for example, all content on a pirated site as illegal content to be monitored. The occurrence prediction unit 14 then stores information related to the specified illegal content in the illegal content collection information storage unit 23, which will be described later. At this time, the information related to the illegal content may be stored in the illegal content collection information storage unit 23 in association with the classification, using classification by the classification unit 16, which will be described later.

Then, the occurrence prediction unit 14 predicts the tendency of illegal content occurrence using, for example, machine learning (AI). The occurrence prediction unit 14 may predict the tendency of illegal content occurrence using, for example, a deep learning LSTM model, which is one of the machine learning models. The occurrence prediction unit 14 may also make predictions using other machine learning models, or may make rule-based predictions. Then, for example, the occurrence prediction unit 14 outputs occurrence prediction data by using as input data the posting data on legitimate content collected by the evaluation collection unit 13 and the occurrence status of illegal content. At that time, the occurrence prediction unit 14 may output prediction data for each classification by using as input data the occurrence status of illegal content for each classification.

The monitoring processing unit 15 monitors the website based on the prediction of illegal content by the occurrence prediction unit 14. When the monitoring processing unit 15 detects the occurrence of illegal content, it requests that the illegal content be removed from the website. The removal request may be made directly to the terminal to which the illegal content was uploaded, but since there is a possibility that the request may not be responded to, it may be reported to an organization that officially manages illegal content, such as a national branch office (management center) that regulates illegal content. Note that the notification method may be, for example, by email.

The classification unit 16 identifies illegal content to be monitored from the additional content determined by the determination unit 12 as not being legitimate content, and classifies the identified illegal content. For example, the classification unit 16 assigns tag information to the illegal content and classifies the content based on the tag information. In this case, for example, the tag information assigned to the legitimate content may be tagged to the illegal content. Furthermore, if tag information is not assigned to the legitimate content, for example, a hashtag (#) obtained by searching an SNS site or the like may be used as tag information to be assigned to the legitimate content. The classification may be, for example, genres such as battle, science fiction, and everyday life, eras such as medieval, modern, and futuristic, and styles such as Japanese and Western. Furthermore, for example, a classification may be obtained from the tag information using a classification database (not shown) that associates tag information with classifications.

The demand forecasting unit 17 forecasts the demand for the content for each classification based on the occurrence status of illegal content for each classification by the classification unit 16 and the number of views and/or downloads of the illegal content.
Here, the demand forecasting unit 17 may forecast the demand for the content for each category for each language. For example, in the case of English, since the language is used in various regions, the demand forecasting unit 17 may further forecast the demand for the content for each category for each region.

The demand prediction unit 17 then predicts the demand for content for each category, for example, using machine learning. The demand prediction unit 17 may predict trends related to the occurrence of illegal content, for example, using an LSTM model. The demand prediction unit 17 may also make predictions using other machine learning models, or may make rule-based predictions. Then, for example, the demand prediction unit 17 outputs demand prediction data by using as input data the posting data for each piece of content collected by the evaluation collection unit 13, and the occurrence status and number of views and/or downloads of illegal content for each category.

The report output unit 18 outputs the deletion request results by the monitoring processing unit 15 as a report, for example, to the customer terminal 4. The report output unit 18 also outputs the demand forecast information by the demand forecasting unit 17 as a report, for example, to the customer terminal 4.

The storage unit 20 is a storage area such as a hard disk, a semiconductor memory device, etc. for storing programs, data, etc. required for the control unit 10 to execute various processes.
The storage unit 20 includes a program storage unit 21 , a legitimate content information storage unit 22 , an illegal content collection information storage unit 23 , and a posted data collection information storage unit 24 .
The program storage unit 21 is a storage area for storing programs. The program storage unit 21 stores various programs including programs for executing various functions of the control unit 10. Note that the program for executing various functions of the control unit 10 may be one program, a different program for each functional unit, or a single program may be configured by multiple functional units.

The legitimate content information storage unit 22 is a storage area for storing information related to legitimate content obtained from a customer. The legitimate content information storage unit 22 stores, for example, the title of the legitimate content, the address (URL) of the website of the legitimate content, etc. The legitimate content information storage unit 22 may also store tag information of the legitimate content.
The illegal content collection information storage unit 23 is a storage area for storing information related to illegal content. The illegal content collection information storage unit 23 stores, for example, the address of the illegal content, the occurrence date and time, the occurrence location, etc. collected by the additional collection unit 11. Here, the occurrence date and time is the period from the date and time of the previous crawling of the website to the date and time of the current crawling of the website and the occurrence of the illegal content is discovered.
The posted data collection information storage unit 24 is a storage area for storing information related to posted data. The posted data collection information storage unit 24 stores, for example, the name of the content included in the posted content, information related to the evaluation of the content, the posting date and time, and the like.

The communication interface unit 29 is an interface for communicating with the customer terminal 4, the web server 7, the SNS server 8, etc. via the communication network N.
Here, a computer refers to an information processing device equipped with a control unit, a memory device, etc., and the collected content analysis server 1, the customer terminal 4, the web server 7, and the SNS server 8 are each information processing devices equipped with a control unit, a memory device, etc., and are included in the concept of a computer.

<Processing Description>
Next, the processing related to the collected content analysis system 100 will be described.
First, the process of collecting illegal contents will be described.
FIG. 2 is a flowchart showing a content collection process in the collected content analysis server 1 according to this embodiment.
This content collection process is always executed while the processing program is running on the collected content analysis server 1 .

In step S (hereinafter, "step S" will be simply referred to as "S") 11 of FIG. 2, the control unit 10 (additional collection unit 11) of the collected content analysis server 1 crawls the websites provided by the web server 7 in order.
In S12, the control unit 10 (additional collection unit 11) determines whether or not additional content has been found by the crawling process. If additional content has been found (S12: YES), the control unit 10 proceeds to S13. On the other hand, if additional content has not been found (S12: NO), the control unit 10 proceeds to S11.

In S13, the control unit 10 (determination unit 12) determines whether the collected additional content is legitimate content or not by referring to the legitimate content information storage unit 22. If the address, title, etc. of the collected additional content match those stored in the legitimate content information storage unit 22, the determination unit 12 can determine that the collected additional content is legitimate content.
In S14, the control unit 10 (discrimination unit 12) judges whether the discrimination result is legitimate content or not. If the discrimination result is legitimate content (S14: YES), the control unit 10 shifts the process to S11. On the other hand, if the discrimination result is not legitimate content (S14: NO), the control unit 10 shifts the process to S15.

In S15, the control unit 10 (generation prediction unit 14) judges whether the additional content is illegal content or not. If it is illegal content (S15: YES), the control unit 10 shifts the process to S16. On the other hand, if it is not illegal content (S15: NO), the control unit 10 shifts the process to S11. Here, the case where it is not illegal content includes the case where it is another legitimate content that is not stored in the legitimate content information storage unit 22.
In S16, the control unit 10 (classification unit 16) classifies the content. To classify the content, the control unit 10 acquires tag information of the illegal content. Then, the control unit 10 determines the classification based on the acquired tag information.
In S17, the control unit 10 (occurrence prediction unit 14) stores the illegal content classification together with the date and time of occurrence and the location of the illegal content in the illegal content collection information storage unit 23. Thereafter, the control unit 10 moves the process to S11. Then, the control unit 10 performs a process of crawling websites in order.

Next, the process of collecting post data will be described.
FIG. 3 is a flowchart showing a posting data collection process in the collected content analysis server 1 according to this embodiment.
Like the content collection process, this post data collection process is always executed while the processing program is running on the collected content analysis server 1 .

In S21 of FIG. 3, the control unit 10 (the evaluation collection unit 13) of the collected content analysis server 1 crawls the SNS sites provided by the SNS server 8 in order.
In S22, the control unit 10 (the evaluation collection unit 13) determines whether or not there is any content posting data by the crawling process. If there is any content posting data (S22: YES), the control unit 10 proceeds to S23. On the other hand, if there is no content posting data (S22: NO), the control unit 10 proceeds to S21.
In S23, the control unit 10 (evaluation collection unit 13) stores the content name, evaluation details, and the date and time and location of the posted data included in the posted data in the posted data collection information storage unit 24. Thereafter, the control unit 10 shifts the process to S21. Then, the control unit 10 performs a process of crawling the SNS sites in order.

Next, a process for predicting and monitoring the occurrence of illegal content will be described.
FIG. 4 is a flowchart showing a content monitoring process in the collected content analysis server 1 according to this embodiment.
FIG. 5 is a diagram showing an example of occurrence prediction processing in the collected content analysis server 1 according to the present embodiment.
The content monitoring process shown in FIG. 4 is executed, for example, once a day at a predetermined time.

In S31 of FIG. 4, the control unit 10 (occurrence prediction unit 14) performs an occurrence prediction process.
Here, the occurrence prediction process will be described.
The occurrence prediction process is a process for predicting trends regarding the occurrence of illegal content, and uses machine learning, for example.
5 shows an example of predicting trends regarding the occurrence of illegal content using a machine learning model, in which occurrence history data 31 and SNS analysis data 32 are input data for the machine learning model, and occurrence prediction data is output. The occurrence history data 31 is generated from data stored in the illegal content collection information storage unit 23, and the SNS analysis data 32 is generated from posting data regarding legitimate content stored in the posting data collection information storage unit 24.

A more specific example will be given. When predicting "when" illegal content will occur by fixing the target content and the location to be monitored, the occurrence history data 31 is time-series data, for example showing the number of illegal content detected each day. The SNS analysis data is also time-series data, for example showing the posting status of the content each day. Here, the posting status is the number of posts given a weight for each post. For example, in the case of a post saying "I want (content name)", the weight coefficient is processed by doubling, for example. In this example, the occurrence prediction data 33, which is output data, is a prediction of the occurrence of illegal content each day.

In S32 of FIG. 4, the control unit 10 (monitoring processing unit 15) monitors the website based on the occurrence prediction.
In S33, the control unit 10 (monitoring processing unit 15) judges whether or not the occurrence of illegal content has been detected. If the occurrence of illegal content has been detected (S33: YES), the control unit 10 proceeds to S34. On the other hand, if the occurrence of illegal content has not been detected (S33: NO), the control unit 10 proceeds to S35.
In S34, the control unit 10 (monitoring processing unit 15) requests, for example, the upload destination of the illegal content to delete the illegal content from the website.

In S35, the control unit 10 judges whether it is time to output the report. In this example, since this process is started once a day, the timing to output the report is, for example, the end of the day (before the next start time). If it is the end of the day and the timing to output the report has arrived (S35: YES), the control unit 10 moves the process to S36. On the other hand, if it is not the timing to output the report (S35: NO), the control unit 10 moves the process to S32.
In S36, the control unit 10 (report output unit 18) performs a report output process of making a report of the deletion request record and outputting it to the customer terminal 4. Thereafter, the control unit 10 ends this process.
The report output timing can be set appropriately. For example, if the report output timing is set to once a month at the end of the month, on days other than the end of the month, even if the result of the process in S35 is NO, this process may be terminated when the start time of the next content monitoring process arrives.

Next, the process of predicting demand for content will be described.
FIG. 6 is a flowchart showing a content demand forecast process in the collected content analysis server 1 according to this embodiment.
The content demand forecast process is executed at an appropriate timing according to, for example, a contract with a customer.

In S51 of FIG. 6, the control unit 10 (demand forecasting unit 17) performs a demand forecast process.
Here, the demand forecasting process will be described.
The demand forecasting process forecasts the demand for content by using the occurrence status and the number of views and/or downloads of illegal content for each language and each category, and the posted data. The demand forecasting for content uses, for example, machine learning.
Demand forecast data, which is the output of machine learning, indicates what categories of content will be popular for each language.

More specifically, for each language and each classification, one of the input data is time-series data, for example, occurrence history data indicating the number of illegal content detections and the number of views (and/or downloads) of the illegal content per day. The other input data is time-series data, for example, SNS analysis data indicating the posting status of content per day. Here, the posting status is the number of posts given a weight for each post, for example, in the case of a post "I want (content name)", the weight coefficient is processed by, for example, doubling. The output data output as a result of the machine learning is a demand forecast indicating the trend of demand for each language and classification.
In S52, the control unit 10 (report output unit 18) performs a report output process of generating a report from the demand forecast data and outputting the report to the customer terminal 4. Thereafter, the control unit 10 ends this process.

As described above, the collected content analysis system 100 of this embodiment has the following advantages.
(1) The collected content analysis server 1 collects content added to a website, and determines whether the added content is legitimate content or not by referring to the legitimate content information storage unit 22. The collected content analysis server 1 then identifies illegal content to be monitored, and predicts the occurrence of illegal content by analyzing the occurrence status of the illegal content.
Therefore, since websites can be monitored based on predictions of illegal content occurrence, it becomes possible to detect the uploading of illegal content at an early stage.
(2) The collected content analysis server 1 classifies illegal contents and predicts the occurrence of illegal contents for each classification.
Therefore, the occurrence of illegal content can be predicted for each category.

(3) The collected content analysis server 1 crawls a specific website.
Therefore, by designating a specific website as a highly illegal website such as a pirated website, it becomes easier to identify illegal content.
(4) The collected content analysis server 1 collects posted data on legitimate content stored in the legitimate content information storage unit 22 from SNS sites, and predicts the occurrence of illegal content using the posted data.
Therefore, by adding posting data to predict the occurrence of illegal content, user opinions are taken into account, leading to more accurate predictions.

(5) The collected content analysis server 1 monitors websites based on predictions of illegal content occurrence, and when it detects the occurrence of illegal content, it requests that the illegal content be removed from the website.
Therefore, by using prediction of illegal content occurrence, it is possible to find illegal content at an earlier stage.
(6) The collected content analysis server 1 outputs the results of the deletion requests to the customer terminal 4 as a report.
Therefore, it is possible to inform customers of the results of removing illegal contents that have been effectively detected using occurrence prediction.

(7) The collected content analysis server 1 classifies illegal contents and predicts demand for contents of each classification based on the occurrence of illegal contents and the number of views (and/or downloads).
Therefore, by analyzing illegal content, it is possible to predict the demand for content that is thought to be related to illegal content.
(8) The collected content analysis server 1 classifies the contents by language and predicts the demand for the contents by language and by classification.
Thus, demand for content can be predicted taking language differences into account.

(9) The collected content analysis server 1 collects posted data about each piece of content from the SNS site, and predicts demand for the content using the posted data.
Therefore, since the demand for the content is predicted by adding the posting data of each content, it becomes possible to perform a more accurate demand prediction.
(10) The collected content analysis server 1 outputs the content demand forecast as a report to the customer terminal 4.
Therefore, a demand forecast for the content can be provided to the customer, and the customer can use the report as a reference to develop a content development strategy.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above-mentioned embodiments. Furthermore, the effects described in the embodiments are merely a list of the most favorable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the embodiments. Note that the above-mentioned embodiments and the modified forms described below can be used in appropriate combinations, but detailed explanations will be omitted.

(Modifications)
(1) In the present embodiment, an example has been described in which illegal content is used in demand forecasting, but this is not limiting. Demand forecasting may also be performed including legitimate content. However, in the case of legitimate content, the customer may make adjustments such as restricting the content. In that case, placing more importance on illegal content is considered to enable a more realistic demand forecast.

(2) In the present embodiment, an example of demand forecasting by language and by category has been described, but the present invention is not limited to this. Demand forecasting may be performed for the content itself, rather than by category. Demand forecasting may also be performed by region, rather than by language, and in this case, the region may be identified using the language.

(3) In this embodiment, an example has been described in which illegal content is identified using information about legitimate content, but this is not limiting. The IP address or the like of a website where illegal content is registered may be registered, and additional content uploaded to the website with that IP address may be identified as illegal content.

(4) In the present embodiment, the posting data is collected by crawling in the same manner as the content is collected, but this is not limiting. For example, in the content monitoring process shown in FIG. 4, the posting data may be collected before the occurrence prediction process (S31).

(5) In the present embodiment, the legitimate content information storage unit is provided in the collected content analysis server, but the present invention is not limited to this. The legitimate content information storage unit may be provided in a device managed by a customer, such as a customer server (not shown), and the collected content analysis server may communicate with the customer's device to refer to the legitimate content information storage unit.

(6) In this embodiment, the SNS site has been described as a site that posts word-of-mouth information, but is not limited to this. The SNS site may be a site such as a blog.

REFERENCE SIGNS LIST 1 collected content analysis server 4 customer terminal 7 web server 8 SNS server 10 control unit 11 additional collection unit 12 discrimination unit 13 evaluation collection unit 14 occurrence prediction unit 15 monitoring processing unit 16 classification unit 17 demand prediction unit 18 report output unit 20 storage unit 21 program storage unit 22 legitimate content information storage unit 23 illegal content collection information storage unit 24 posted data collection information storage unit 100 collected content analysis system

Claims (9)

an additional collection means for collecting additional content added to a website from a website where the content can be viewed and/or downloaded;
a legitimate content determination unit that determines whether the additional content collected by the additional collection unit is a legitimate content by referring to a legitimate content information storage unit that stores information that can identify the legitimate content to be monitored;
a first evaluation collection means for collecting first posting data on the legitimate content stored in the legitimate content information storage unit from an SNS site;
an occurrence prediction means for identifying illegal content to be monitored that is included in the additional content that has been determined to be non-authorized content by the authorized content discrimination means, and obtaining occurrence prediction data using a machine learning model trained using an LSTM model that outputs the occurrence of the illegal content per day, using occurrence history data that is time series data on the number of detections per day generated from the occurrence date and time of the identified illegal content and time series data on the number of posts per day generated from the occurrence date and time of the first posting data collected by the first evaluation collection means as inputs ;
A collection content analysis device comprising: 2. The collected content analysis apparatus according to claim 1,
a content classification means for classifying the illegal content included in the additional content determined as not being the legitimate content by the legitimate content determination means,
The occurrence prediction means receives the occurrence history data for each classification by the content classification means as an input, and obtains the occurrence prediction data for each classification using the machine learning model. 3. The collected content analysis device according to claim 1,
The additional collection means collects the additional content by crawling the specific website. 4. The collected content analysis device according to claim 1,
a monitoring means for monitoring the website based on the occurrence prediction data obtained by the occurrence prediction means;
a deletion request means for requesting the deletion of the illegal content from the website when the occurrence of the illegal content is detected by the monitoring means;
A collection content analysis device comprising: 5. The collected content analysis apparatus according to claim 4 ,
The collected content analysis device further comprises a request record output means for outputting deletion request record information made by the deletion request means. 6. The collected content analysis device according to claim 1,
a content classification means for classifying the illegal content included in the additional content determined as not being the legitimate content by the legitimate content determination means;
a second evaluation collection means for collecting second posting data for each piece of content from the SNS site;
a demand forecasting means for obtaining demand forecast data using a machine learning model trained using an LSTM model with inputs of occurrence history data, which is time series data of the number of detections and the number of views and/or downloads per day generated based on the occurrence date and time and the number of views and/or downloads of the illegal content for each classification by the content classification means, and time series data of the number of posts per day generated from the occurrence date and time of the second posting data collected by the second evaluation collection means, so that the demand for the content for each classification is an output;
A collection content analysis device comprising: 7. The collected content analysis apparatus according to claim 6 ,
The content classification means classifies the content by language,
The demand prediction means receives the generation history data for each language from the content classification means as an input, and obtains the demand prediction data for each language using the machine learning model. 8. The collected content analysis device according to claim 6 ,
A collected content analysis device comprising a demand forecast output means for outputting the demand forecast data obtained by the demand forecast means in the form of a report. A program for causing a computer to function as the collected content analysis device according to any one of claims 1 to 8 .

Images