JP2011166621A - Video-content recommendation apparatus, method for determining recommended video content, and computer program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To specify a new or topical video content so as to recommend the video content. <P>SOLUTION: A video-content recommendation apparatus includes: a data collecting part 11 for collecting text data from websites and video sites on the Internet; a keyword extracting part 13 for extracting video-content-related keywords, being keywords related to a plurality of video contents, from the collected text data; and a recommended-content determining part 17 that makes a ranking of the plurality of video contents on the basis of the extracted video-content-related keywords so as to determine a recommended video content from among the plurality of video contents on the basis of the ranking. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a technology for identifying new or topical video content and recommending the video content to a user.

  Conventionally, various techniques for recommending video content such as television programs to users have been proposed. For example, in Patent Document 1, a user's preference is estimated based on an action history such as a viewing history of video content by a user and an access history to a website, and video content is recommended according to the preference. Are listed.

JP 2004-343321 A

  Like the technique described in Patent Document 1, the fact that recommended content is determined based on a user's past action history means that video content is recommended based on a user's actualized preference. It is. Therefore, the content recommendation based on such a manifested preference does not give the user a new encounter or unexpectedness.

  On the other hand, many users have a desire to quickly obtain information on new or topical video content. On the other hand, in the technique of Patent Document 1, the user's preference genre is set and the recommended content is determined based on this, but this is not always sufficient.

  Therefore, an object of the present invention is to identify new or topical video content and recommend the video content.

  An apparatus for recommending video content according to an embodiment of the present invention relates to a collection unit that collects text data from websites and video sites on the Internet, and a plurality of video contents from the text data collected by the collection unit. Extraction means for extracting video content related keywords as keywords, ranking means for ranking the plurality of video contents based on the video content related keywords extracted by the extraction means, and ranking by the ranking processing unit And recommending means for determining a recommended video content from among the plurality of video content.

  In a preferred embodiment, in the text data collected by the collecting means, an appearance frequency calculating means for calculating an appearance frequency for each video content related keyword extracted by the extracting means, and each of the plurality of video contents Storage means for storing the content reference information according to the above. The ranking unit associates the video content related keyword with the plurality of video content based on the content reference information stored in the storage unit, and according to the appearance frequency for each video content related keyword, You may make it perform ranking of several video content.

  In a preferred embodiment, in the text data collected by the collecting means, an appearance frequency calculating means for calculating an appearance frequency for each video content related keyword extracted by the extracting means, and each of the plurality of video contents The storage means for storing the content reference information related to the above, and the latest appearance frequency and the past appearance frequency among the appearance frequencies for each video content related keyword calculated by the calculation means are compared, and the appearance for each video content related keyword An increase rate calculating means for calculating an increase rate of the frequency may be further provided. The ranking unit associates the video content related keyword with the plurality of video content based on the content reference information stored in the storage unit, and responds to an increase in appearance frequency for each video content related keyword. The ranking of the plurality of video contents may be performed.

1 is an overall configuration diagram of a video content recommendation system according to an embodiment of the present invention. 1 is a configuration diagram of a server 1. FIG. An example of the data structure of the extraction keyword memory | storage part 33 is shown. It is explanatory drawing of the calculation procedure of an increase rate. An example of the data structure of the content reference information storage part 35 is shown. It is explanatory drawing which shows the specific example of the ranking process of a video content. It is a flowchart which shows the process sequence of the recommendation content determination by the total appearance frequency of an extraction keyword. It is a flowchart which shows the process sequence of the recommendation content determination by the maximum increase rate of an extraction keyword.

  Hereinafter, a video content recommendation system according to an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is an overall configuration diagram of a video content recommendation system according to an embodiment of the present invention.
As shown in the figure, the present system includes a server 1 that performs processing related to recommendation of video content such as a television program (hereinafter sometimes simply referred to as “content”). A server 1 is connected to a user terminal device 3 and an STB (set top box) 4 used by a user via a network 9 such as the Internet. Further, the server 1 is connected to websites 7, 7,... On the Internet and moving image sites 8, 8,.

  The server 1 is configured by, for example, a general-purpose computer system, and individual components or functions in the server 1 described below are realized by, for example, executing a computer program. This computer program can be stored in a computer-readable recording medium.

  The user terminal device 3 may be a terminal device having a communication function, and may be, for example, a mobile phone, a portable information terminal, or a general-purpose personal computer. The user terminal device 3 includes a display device such as a liquid crystal panel, and an input device such as a push button, a pointing device, or a touch panel. When the user terminal device 3 is a mobile phone or the like, wireless communication is performed with a wireless base station.

  The STB 4 receives and records video content (stores in a storage device) in accordance with an instruction from the server 1 or an instruction from the user. Furthermore, the STB 4 outputs the received or recorded video content to a television receiver (not shown) according to an instruction from the server 1 or an instruction from the user, and reproduces the video content. In this embodiment, an STB is described as an example of a recording apparatus. However, in addition to the STB, for example, an HDD recorder, a television receiver with a recording function (including a one-segment receiver), a PC with a recording function, and the like Good.

  FIG. 2 shows a configuration diagram of the server 1 as a video content recommendation device.

  The server 1 includes a data collection unit 11, a keyword extraction unit 13, an increase rate determination unit 15, a recommended content determination unit 17, a recording reservation processing unit 19, a recommended mail generation unit 21, and a content related keyword storage unit 31. And an extracted keyword storage unit 33 and a content reference information storage unit 35.

  The data collection unit 11 crawls data from various sites on the Internet. For example, the data collection unit 11 collects text data from the website 7 and the moving image site 8. The data collected by the data collection unit 11 is, for example, information transmitted by an unspecified number of people on blogs and SNSs (social networking services), or specific people (organizations) such as celebrities and mass media. Information. Such information includes highly topical or novel information.

  The data collection unit 11 can collect data at an arbitrary timing. For example, in the following embodiment, data collection is performed once a day.

  The content-related keyword storage unit 31 stores content-related keywords that are keywords related to a plurality of video contents. A content-related keyword predetermined by the administrator of the system may be registered in the content-related keyword storage unit 31. Alternatively, the content-related keyword may be extracted from content reference information stored in the content reference information storage unit 35. Alternatively, the content-related keyword may be extracted from a search ranking or the like published by a search site or a portal site.

  The keyword extraction unit 13 extracts content-related keywords from the text data collected by the data collection unit 11. For example, the keyword extraction unit 13 performs morphological analysis on the text data collected by the data collection unit 11 and breaks it down into words. The keyword extraction unit 13 extracts words that match the content-related keywords stored in the content-related keyword storage unit 31 from the words acquired here. Here, the word (content-related keyword) extracted by the keyword extraction unit 13 is referred to as an extracted keyword.

  The keyword extraction unit 13 further calculates the appearance frequency for each extracted keyword extracted here in the text data collected by the data collection unit 11. For example, the keyword extraction unit 13 may count the number of appearances of each extracted keyword included in the text data collected by the data collection unit 11. The keyword extraction unit 13 may count the number of appearances at the same time as the above-described word cutout, and then extract content-related keywords.

  The keyword extraction unit 13 associates the extracted keywords extracted here with their appearance frequencies and stores them in the extracted keyword storage unit 33.

  The extracted keyword storage unit 33 stores the keywords extracted by the keyword extraction unit 13.

  FIG. 3 shows an example of the data structure of the extracted keyword storage unit 33. As shown in the figure, the extracted keyword storage unit 33 has an extraction date and time 331, an extracted keyword 333, an appearance count 335, and an increase rate 337 as data items.

  The extraction date and time 331 indicates the date and time when the data collection unit 11 performs data collection and the extracted keyword is extracted.

  The extracted keyword 333 is a keyword extracted by the keyword extracting unit 13.

  The number of appearances 335 is the number of appearances of the extracted keyword 333 counted by the keyword extraction unit 13.

  The increase rate 337 is an increase rate with respect to the appearance count 335 at the previous extraction. The increase rate 337 is calculated by the increase rate determination unit 15 described below.

  Referring to FIG. 2 again, the increase rate determination unit 15 refers to the extracted keyword storage unit 33, compares the latest appearance frequency with the past appearance frequency among the appearance frequencies for each extracted keyword, and extracts each keyword. The rate of increase in the appearance frequency of is calculated.

  For example, the increase rate determination unit 15 compares the number of appearances 335 of a certain extracted keyword 333 with the number of appearances 335 of the same extracted keyword 333 in the past (for example, at the time of previous extraction), and increases the rate of increase from the past number of appearances 335. Is calculated.

  The procedure for calculating the increase rate will be specifically described with reference to FIG.

  First, keywords extracted from data collected during the previous processing of this processing (January 1, 2010, 21:15:45) and the number of appearances thereof are as shown in FIG. 4A. For the same keywords as the extracted keywords shown here, the number of appearances at the time of the current processing (January 2, 2010, 10:02:32) is as shown in FIG. 4B. At this time, when the increase rate determination unit 15 calculates the increase rate using the appearance count at the previous processing as the denominator and the appearance count at the current processing as the numerator, the increase rate is as shown in FIG. 4C. The increase rate calculated here is stored in the increase rate 337 at the time of the current process.

  In addition, the increase rate determination unit 15 increases the appearance frequency for each extracted keyword based on the appearance frequency for each extracted keyword extracted from text data related to the blog, microblog, SNS, and the like collected by the data collection unit 11. May be calculated. This is because the content described in blogs, microblogs, SNSs, and the like has many descriptions especially for new ones and is considered to reflect the fashion in a sensitive manner.

  Referring to FIG. 2 again, the content reference information storage unit 35 stores content reference information related to each video content for a plurality of video content. The content reference information is information related to each video content, and may be, for example, EPG (Electric Program Guide) program information, video content information of a video site, and the like.

  FIG. 5 shows an example of the data structure of the content reference information storage unit 35. As shown in the figure, the content reference information storage unit 35 includes a content title 351, a broadcast date 352, a time zone 353, a broadcast station 354, a genre 355, a content overview 356, a content detail 357, and an associated person name 358. Is included as a data item. The related person name 358 may be, for example, a performer of the content. In addition, the content reference information storage unit 35 may include comments and evaluations related to the program.

  Returning to FIG. 2, the recommended content determination unit 17 ranks a plurality of video contents based on the extracted keywords. Further, the recommended content determination unit 17 determines a recommended video content from a plurality of video content based on the ranking.

  For example, the recommended content determination unit 17 associates the extracted keyword with a plurality of video contents based on the content reference information stored in the content reference information storage unit 35, and determines a plurality of items according to the number of appearances for each extracted keyword. Ranking video content. For example, when the extracted keyword is included in the text included in the content outline 356, the content details 357, the person name 358, and the like, the recommended content determination unit 17 associates the video content with the extracted keyword 333. The recommended content determination unit 17 sums up the appearance frequency 335 of the extracted keyword 333 associated with each content. The recommended content determination unit 17 ranks the video content in descending order of the total number of appearances.

  The recommended content determination unit 17 associates the extracted keyword with a plurality of video contents based on the content reference information stored in the content reference information storage unit 35, for example, and increases the appearance frequency for each extracted keyword. Accordingly, a plurality of video contents are ranked. For example, if the extracted keyword is included in the text included in the content summary 356, the content details 357, the person name 358, and the like, the recommended content determination unit 17 determines the video content and the extracted keyword 333 as described above. Associate. The recommended content determination unit 17 specifies the largest numerical value (maximum increase rate) among the increase rates 337 of the extracted keywords 333 associated with each content. The recommended content determination unit 17 ranks the video content in descending order of the maximum increase rate.

  FIG. 6 is an explanatory diagram showing a specific example of ranking processing of video content.

First, extracted keywords (“A”, “B”, “C”, “D”) are stored in the extracted keyword storage unit 33 together with the number of appearances and the rate of increase. Here, when the recommended content determination unit 17 associates the content reference information in the content reference information storage unit 35 with the extracted keyword,
Content “A” and extracted keyword “A”,
Content “I” and extracted keywords “B” “C”
Correspondence relationships between the content “U” and the extracted keyword “D” and the content “E” and the extracted keyword “A” are obtained.

  When the recommended content determination unit 17 calculates the total number of appearances of each content, the content “I” is 70, the content “A” is 50, the content “E” is 50, and the content “U” is 20. Therefore, the recommended content determination unit 17 determines the content ranking 210 based on the total number of appearances. If the total number of appearances is the same, content with an earlier broadcast date and time may be ranked higher.

  On the other hand, when the recommended content determination unit 17 calculates the maximum increase rate of each content, the content “U” is 1.4, the content “I” is 1.2, the content “A” is 0.8, and the content “E” Is 0.8. Therefore, the recommended content determination unit 17 determines the content ranking 220 based on the maximum increase rate. As for the maximum increase rate, if the rate is the same, content with an earlier broadcast date and time may be ranked higher.

  The recording reservation processing unit 19 instructs the STB 4 to make a recording reservation for the recommended content determined by the recommended content determination unit 17.

  The recommended mail generation unit 21 transmits the recommended content determined by the recommended content determination unit 17 to the user terminal device 3 as a recommended email for notifying the recommended content. When the recommended e-mail received by the user terminal device 3 is displayed, the recommended content determined by the recommended content determination unit 17 is displayed.

  In the present embodiment, the recommended content is notified to the user terminal device 3 using e-mail, but in addition to this, for example, the Web browser of the user terminal device 3 makes a request to the server 1 to make a recommendation. You may make it acquire a content.

  Next, a processing procedure for determining recommended content in the above-described video content recommendation system will be described with reference to the flowcharts of FIGS.

  FIG. 7 is a flowchart illustrating a processing procedure for determining recommended content based on the total number of appearances of extracted keywords.

  First, the data collection unit 11 collects text data from the website 7 and the moving image site 8 on the Internet (S100). The keyword extraction unit 13 cuts out words from the text data collected in step S100 by morphological analysis or the like, and counts the number of appearances of each word (S110). The recommended content determination unit 17 extracts words that match the content-related keywords stored in the content-related keyword storage unit 31 as extracted keywords (S120). The extracted keywords extracted here and the number of appearances thereof are stored in the extracted keyword storage unit 33.

  The recommended content determination unit 17 refers to the extracted keyword storage unit 33 and the content reference information storage unit 35, and associates each content with the extracted keyword (S130). Further, the recommended content determination unit 17 calculates the total number of appearances based on the number of appearances of each associated extracted keyword, and ranks each content (S140). The recommended content determination unit 17 determines recommended content based on the ranking result (S150).

  The recording reservation processing unit 19 gives a recording reservation instruction based on the recommended content determined in step S150, and the recommended email generation unit 21 transmits a content recommendation email based on the recommended content determined in step S150 (S160). . When the recommended e-mail received by the user terminal device 3 is displayed, the recommended content determined by the recommended content determination unit 17 is displayed.

  Thereby, it is possible to promptly recommend to a user video content related to a keyword that has been written a lot on the Internet.

  FIG. 8 is a flowchart showing a procedure for determining recommended content based on the maximum increase rate of extracted keywords. Processes common to the flowchart of FIG. 7 are given the same step numbers.

  First, the processing from step S100 to S120 by the data collection unit 11 and the keyword extraction unit 13 is the same as in the case of FIG.

  The increase rate determination unit 15 refers to the extracted keyword storage unit 33, and calculates an increase rate of the number of appearances at the time of current extraction with respect to the number of appearances at the time of previous extraction of each extracted keyword (S200).

  The recommended content determination unit 17 refers to the extracted keyword storage unit 33 and the content reference information storage unit 35, and associates each content with the extracted keyword (S130). Further, the recommended content determination unit 17 calculates the highest increase rate for each content based on the increase rate of each extracted keyword associated with each other, and ranks each content (S210). The recommended content determination unit 17 determines recommended content based on the ranking result (S150).

  The recording reservation processing unit 19 and the recommended mail generation unit 21 transmit a recording reservation instruction and a recommended mail as in the case of FIG. 7 (S160). When the recommended e-mail received by the user terminal device 3 is displayed, the recommended content determined by the recommended content determination unit 17 is displayed.

  As a result, it is possible to promptly recommend to the user video content related to a new keyword whose number of times of suddenly rising on the Internet.

  The above-described embodiments of the present invention are examples for explaining the present invention, and are not intended to limit the scope of the present invention only to those embodiments. Those skilled in the art can implement the present invention in various other modes without departing from the gist of the present invention.

1 Server 3 User terminal device 7 Website 8 Video content 11 Data collection unit 13 Keyword extraction unit 15 Increase rate determination unit 17 Recommended content determination unit 19 Recording reservation processing unit 21 Recommended mail generation unit 31 Content related keyword storage unit 33 Extracted keyword storage Unit 35 Content reference information storage unit

Claims (5)

A collection means for collecting text data from websites and video sites on the Internet;
Extraction means for extracting video content-related keywords, which are keywords related to a plurality of video contents, from the text data collected by the collecting means;
Ranking means for ranking the plurality of video contents based on the video content-related keywords extracted by the extraction means;
A video content recommendation device comprising: recommendation means for determining recommended video content from among the plurality of video content based on ranking by the ranking processing unit. Appearance frequency calculating means for calculating the appearance frequency for each video content related keyword extracted by the extracting means in the text data collected by the collecting means;
Storage means for storing content reference information related to each of the plurality of video contents,
The ranking means associates the video content-related keyword with the plurality of video contents based on the content reference information stored in the storage means, and determines the plurality of video content-related keywords according to the appearance frequency for each video content-related keyword. The content recommendation device according to claim 1, wherein ranking of video content is performed. Appearance frequency calculating means for calculating the appearance frequency for each video content related keyword extracted by the extracting means in the text data collected by the collecting means;
Storage means for storing content reference information related to each video content for the plurality of video content;
An increase rate calculating means for comparing the latest appearance frequency with the past appearance frequency among the appearance frequencies for each video content related keyword calculated by the calculating means, and calculating an increase rate of the appearance frequency for each video content related keyword; Further comprising
The ranking means associates the video content related keyword with the plurality of video contents based on the content reference information stored in the storage means, and according to the degree of increase in appearance frequency for each video content related keyword, The content recommendation device according to claim 1, wherein ranking of the plurality of video contents is performed. Collecting text data from websites and video sites on the Internet;
Extracting video content related keywords, which are keywords related to a plurality of video content, from the collected text data;
Ranking the plurality of video contents based on the extracted video content-related keywords;
Determining a recommended video content from the plurality of video content based on the ranking, and determining a recommended video content. A computer program for determining recommended video content,
When executed on a computer,
Collecting text data from websites and video sites on the Internet;
Extracting video content related keywords, which are keywords related to a plurality of video content, from the collected text data;
Ranking the plurality of video contents based on the extracted video content-related keywords;
Determining a recommended video content from the plurality of video content based on the ranking.

Images