JP2015158858A - Information providing device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information providing device and program that allow information to be provided in accordance with a status of a user on the basis of a usage of a publicly opened SNS.SOLUTION: An information providing device 10 comprises: a collection unit 12 that collects utterances in a social networking group service of a user; an extraction unit 13 that extracts the utterance about a broadcast program or broadcasting station from the utterances collected by the collection unit 12; a first trend analysis unit 14 that analyzes a trend of a time zone in which the user views the broadcast program on the basis of the utterance extracted by the extraction unit 13; a second trend analysis unit 15 that analyzes a trend of the broadcast program viewed by the user on the basis of the utterance extracted by the extraction unit 13; a classification unit 16 that classifies the user in accordance with a viewing trend on the basis of the analysis result by the first trend analysis unit 14 and the analysis result by the second trend analysis unit 15; and a providing unit 17 that provides information to the user in accordance with a result of the classification obtained by the classification unit 16.

Description

  The present invention relates to an information presentation apparatus and a program for presenting information to a user based on analyzing a user's program viewing tendency using a statement in a social networking service (hereinafter referred to as “SNS”).

  Conventionally, there is a method of presenting information according to a user's situation. For example, the first method is a method of recommending a restaurant based on the situation such as time, weather, budget, accompanying persons, number of people (see Non-Patent Document 1). The second method is a method of recommending an application in consideration of the application usage history of the mobile terminal and the user's situation (see Non-Patent Document 2).

Kenta Oku, “Study on Information Recommendation Method Considering User Context”, Nara Institute of Science and Technology, Doctoral Dissertation, 2009 ) Jun Osawa, “Analysis of relationship between context and application used to realize mobile application recommendation system”, Osaka University, Graduation Research Report, 2012 (http://www-nishio.ist.osaka-u.ac. jp / Thesis / bachelor / 2011 / osawa / thesis.pdf)

  In the non-patent documents 1 and 2 described above, since the user's behavior history, operation history, and the like are used, it is necessary to clear the problem of personal information protection. In addition, information that cannot be automatically acquired from the mobile terminal must be directly input by the user. For this reason, in order to have the user recommend information, it is necessary to consciously perform an input operation.

  An object of this invention is to provide the information presentation apparatus and program which can provide the information according to a user's state based on using SNS publicly disclosed.

  The information presentation device includes a collection unit that collects comments in a social networking service of a user who has an account in a social networking service, an extraction unit that extracts a broadcast program or a broadcast station comment from the collected messages by the collection unit, A first trend analysis unit that analyzes a trend of a time zone during which a user views a broadcast program based on the speech extracted by the extraction unit, and a broadcast program that the user views based on the speech extracted by the extraction unit A second trend analysis unit that analyzes the trend, a classification unit that classifies users according to viewing trends based on the analysis result of the first trend analysis unit and the analysis result of the second trend analysis unit; A presentation unit that presents information to the user in accordance with the classification result obtained by the classification unit.

  In addition, the first tendency analysis unit obtains the utterance time zone of the utterance and the number of utterances corresponding to the utterance time zone from the utterance extracted by the extraction unit, so that the user can watch the broadcast program. It is preferable to analyze the trend of the time zone.

  In addition, when the utterance extracted by the extraction unit includes a broadcast program name, the second tendency analysis unit identifies the utterance regarding the broadcast program with the broadcast program name, and the utterance extracted by the extraction unit If the broadcast program name is not included in the message, the message is specified as a message related to the broadcast program broadcast in the message time zone, and is obtained based on the broadcast program corresponding to the specified message and the specified message. It is preferable to analyze the tendency of the broadcast program viewed by the user based on the number of utterances for each broadcast program.

  The information presentation device preferably includes a registration unit that registers data of an account of a social networking service used by the user. In this case, the collection unit collects comments in a social networking service of a user who has an account registered by the registration unit.

  The program is executed on a computer connected to a communication network. The program includes a collecting step of collecting a utterance in a social networking service of a user who has an account in the social networking service, and an extraction step of extracting a utterance including a broadcast program name or a broadcasting station name from the utterance collected by the collecting step; A first trend analysis step for analyzing a trend of a time zone in which the user views a broadcast program based on the speech extracted by the extraction step, and a broadcast program that the user views based on the speech extracted by the extraction step A classification step for classifying users according to viewing trends based on a second trend analysis step for analyzing a trend, an analysis result of the first trend analysis step, and an analysis result of the second trend analysis step; Present information to the user according to the classification result And the presenting step that, to run.

  ADVANTAGE OF THE INVENTION According to this invention, the information presentation apparatus and program which can provide the information according to a user's state can be provided based on utilizing SNS publicly disclosed.

It is a figure for demonstrating one Embodiment of the information presentation system containing an information presentation apparatus. It is a figure for demonstrating the example of the broadcast program recommended to a user.

  Hereinafter, embodiments of the present invention will be described. FIG. 1 is a diagram for explaining an embodiment of an information presentation system 1 including an information presentation device 10.

The information presentation system 1 includes an information presentation device 10, a user terminal 21, a web server 22, an EPG (Electronic Program Guide) server 23, and a communication network NW.
One or a plurality of user terminals 21 are provided. The user has an account in the SNS on the web server 22 based on using the user terminal 21. A specific example of SNS is Twitter. The user terminal 21 is, for example, a smartphone or a tablet computer.
The EPG server 23 has data related to EPG information of a program broadcast from a broadcast station (not shown).

  The information presentation device 10 includes a registration unit 11, a collection unit 12, an extraction unit 13, a first trend analysis unit 14, a second trend analysis unit 15, a classification unit 16, and a presentation unit 17.

  The registration unit 11 registers SNS account data used by the user. The user is a user who wants to provide information from the information presentation device 10. The SNS account data is, for example, a URL (Uniform Resource Locator), an account name, or an ID. That is, the registration unit 11 registers a page on which a user's remark is displayed. Thereby, the information presentation apparatus 10 can access the page of the user who desires to provide information.

  The collection unit 12 collects comments in a social networking service of a user who has an account in the SNS. That is, the utterances (writes) in the SNS of the user who has the account registered by the registration unit 11 are collected. Specifically, the collection unit 12 collects a user's remarks (writing) using an API (Application Programming Interface) provided from the SNS side.

  The extraction unit 13 extracts a comment related to a broadcast program or a broadcast station from the comments collected by the collection unit 12. For example, the extraction unit 13 extracts a comment with a tag indicating a broadcast program name or a broadcast station name. As a more specific example, if the SNS is Twitter, the extraction unit 13 extracts a message with a hash tag indicating a broadcast program name or a broadcast station name. For example, if the utterance includes a broadcast program name or a broadcast station, the extraction unit 13 extracts the utterance. Further, the extraction unit 13 compares the EPG information of the EPG server 23 with the SNS remarks, and in the broadcast time zone of a broadcast program, the name or title of the performer of the broadcast program, the subtitle of the broadcast program, or If a comment including the program content of the broadcast program is posted, it is extracted as a comment about the broadcast program.

  The first tendency analysis unit 14 analyzes the tendency of the time zone when the user views the broadcast program based on the remarks extracted by the extraction unit 13. That is, the first tendency analysis unit 14 obtains the utterance time zone of the utterance and the number of utterances corresponding to the utterance time zone from the utterance extracted by the extraction unit 13, so that the user views the broadcast program. Analyze time zone trends.

More specifically, the first trend analysis unit 14 first counts the number of extracted utterances. That is, the 1st tendency analysis part 14 totals the number of utterances for every time of a certain day. As an example, the first tendency analysis unit 14 divides a time range from 0 o'clock to 23 o'clock every hour, and counts the number of utterances per hour. The first trend analysis unit 14 may divide every two hours instead of every hour, or may be divided every morning (4:00 to 10:00), noon (10: 0 to 18:00), You may divide like night (18:00 to 23:00) and midnight (23:00 to 4:00). In addition, the utterance regarding the broadcast program to the SNS may slightly deviate from the broadcast time of the broadcast program. For this reason, the 1st tendency analysis part 14 seems to overlap in time zones, such as (1: 00-2: 30) * (2: 00-3: 30) * (3: 00: 00-4: 30). May be separated into
However, the separation method used for one trend analysis is all unified in the same separation method.

And the 1st tendency analysis part 14 produces the vector shown to the following (1) Formula as a viewing-and-listening tendency.
A _{u, d} = <(t ₀ , a ₀ ), (t ₁ , a ₁ ), ..., (t ₂₃ , a ₂₃ )> (1)
Here, (1) vector of expression hours in d∈D day in user _{u t n ∈ {t 0,} ..., t 23} are represented by the write number _{a n} of each. The vector of equation (1) is obtained by dividing the hour from 0 o'clock to 23 o'clock every hour, and totaling the number of utterances per hour.

  The first trend analysis unit 14 creates vectors for a plurality of days. The first tendency analysis unit 14 calculates the inner product value for any two of the created vectors. As an example, the first trend analysis unit 14 calculates inner product values for two vectors related to the same day of the week. As a result of the calculation, the first tendency analysis unit 14 has a similar time every day as the inner product value is higher (for example, higher than the threshold) in more combinations (for example, at a certain ratio or more in the combination). It is determined that the program is being watched on the band.

  The second tendency analysis unit 15 analyzes the tendency of the broadcast program that the user views based on the remarks extracted by the extraction unit 13. That is, when the broadcast program name is included in the utterance extracted by the extraction unit 13, the second tendency analysis unit 15 specifies the utterance related to the broadcast program with the broadcast program name. On the other hand, when the broadcast program name is not included in the utterance extracted by the extraction unit 13, the second tendency analysis unit 15 specifies the utterance regarding the broadcast program broadcast in the utterance time zone of the utterance. And the 2nd tendency analysis part 15 analyzes the tendency of the broadcast program which a user views based on the broadcast program corresponding to the specified message, and the number of messages for every broadcast program obtained based on the specified message. .

  More specifically, the second trend analysis unit 15 first analogizes which broadcast program the utterance is for. When the tag indicating the broadcast program name is attached to the utterance, the second tendency analysis unit 15 assumes that the utterance is for the broadcast program with the broadcast program name. In addition, when a comment including the name of the performer of the broadcast program is posted in the broadcast time zone of a certain broadcast program, the second tendency analysis unit 15 analogizes that the comment is about the broadcast program. . In addition, the second tendency analysis unit 15 analyzes the utterance extracted by the extraction unit 13 (for example, morphological analysis), and based on the utterance content and the utterance time that are the results of the analysis, the second trend analysis unit 15 generates an EPG of the EPG server 23. Infer by referring to information.

Next, the 2nd tendency analysis part 15 totals the number of utterances for every program or every genre of a program. For example, the second tendency analysis unit 15 creates vectors shown in the following formulas (2) and (3).
B _u = <(p ₀ , b ₀ ), (p ₁ , b ₁ ),..., (P _n , b _n )> (2)
C _s = <(p ₀ , e ₀ ), (p ₁ , e ₁ ),..., (P _n , e _n )> (3)
Here, the vector of the equation (2) is represented by the number of writes b _m for the broadcast program p _m εP = {p ₀ ,..., P _n } of the user u. The vector of equation (3) represents a vector of a certain program series S⊂P. In _addition, e 0, ... each of e _n, next to 1 if the program of the program series S, becomes 0 if it is not the program of the program series S.

  Next, the second trend analysis unit 15 calculates the inner product value of the vector B representing the number of user writes and the vector C representing the series program. As a result of the calculation, the second trend analysis unit 15 determines that the user is paying more attention to the series program as the inner product value is higher (for example, higher than the threshold value).

  The classification unit 16 classifies the users according to the viewing tendency based on the analysis result of the first trend analysis unit 14 and the analysis result of the second trend analysis unit 15. That is, for example, the classification unit 16 views the same program every week for a user who watches at a fixed time every day, depending on whether the user's viewing time zone is fixed or the user's viewing program is fixed. They are classified as a user or a user who watches the same program with no fixed time zone.

  The presenting unit 17 presents information to the user according to the classification result obtained by the classifying unit 16. The information presented to the user is, for example, a broadcast program recommended for the user, a video recording medium, an audio recording medium, or a book.

FIG. 2 is a diagram for explaining an example of a broadcast program recommended to the user.
As an example, when the classification unit 16 classifies the user as watching a fixed program (same program) at the same time zone every day, the user is presented as viewing a specific program as a daily routine. The unit 17 recommends another program similar to the specific program in the vicinity of the daily viewing time to the user, or a related information page or VOD (Video On Demand) that can further deepen the daily viewing program. ) Etc.

  As an example, if the classification unit 16 classifies the user as a user who is watching a program irregularly in the same time zone on an irregular day (when the program being viewed is not fixed), the life rhythm The presentation unit 17 recommends a recommended program set in advance near the viewing time zone to the user, or a friend on the SNS views it. Recommend programs derived by other methods, such as programs that are currently running.

  In FIG. 2, when the time is “daily” and “irregular” and the program is “fixed”, “recommend another program based on content” is to create a profile from the user's viewing history, The recommendation is to recommend a program that is similar to a program you often watch. The recommended method will be described below.

First, the presentation unit 17 creates a program profile. The presentation unit 17 assigns an evaluation value to each word included in the EPG information for each broadcast program by performing morphological analysis on the EPG information of the EPG server 23, and provides the characteristics of the program d _i represented by the following equation (4). This is represented by a vector (program vector) shown.
D _i = <w _i1 ,..., W _im > (4)
Here, w _ij is the weight of the word k _j in the program d _i . w _ij is expressed by the following equation (5).
w _ij = e _ij · idf _j (5)
Here, e _ij is 1 if the word k _j is included in the program d _i , and 0 if it is not included. idf _j represents the rarity of the word k _j (Inverse Document Frequency).

…（６）
ここで、ｂ_ｍは、番組ｄ_ｍの希少性を表す。 Next, the presentation unit 17 creates a user profile for the user. The presentation unit 17 adds the program vectors of the programs viewed by the user, and creates a profile vector of the user _qr shown in the following equation (6).

... (6)
Here, _{b m} represents the scarcity of programs _{d m.}

  Next, the presentation unit 17 selects a broadcast program (content) having a high degree of similarity to the broadcast program viewed by the user based on the program vector shown in Equation (4) and the profile vector shown in Equation (6). Recommendation to. A broadcast program with a high degree of similarity is a vector as shown in Expression (4) based on EPG information of a broadcast program to be broadcast from now on, and the inner product value of the vector and the vector shown in Expression (6) It is a broadcast program that shows a high value. The presentation unit 17 recommends the broadcast program.

  As described above, the information presentation apparatus 10 includes a collection unit 12 that collects utterances in the user's SNS, an extraction unit 13 that extracts utterances related to broadcast programs or broadcast stations from the utterances collected by the collection unit 12, and an extraction unit 13. The first tendency analysis unit 14 that analyzes the trend of the time zone during which the user views the broadcast program based on the utterance extracted by the user, and the broadcast program that the user views based on the utterance extracted by the extraction unit 13 A second trend analysis unit 15 that analyzes a trend, a classification unit 16 that classifies users according to viewing tendencies, based on the analysis result of the first trend analysis unit 14 and the analysis result of the second trend analysis unit 15; And a presentation unit 17 that presents information to the user in accordance with the classification result obtained by the classification unit 16. Thereby, the information presentation apparatus 10 can provide the information according to a user's state (classification | category) based on utilizing SNS publicly disclosed.

  The first tendency analysis unit 14 obtains the speech time zone of the speech and the number of speeches corresponding to the speech time zone from the speech extracted by the extraction unit 13, so that the user can watch the broadcast program Analyzing trends. Thereby, the information presentation apparatus 10 can obtain a viewing tendency of a broadcast program, particularly a tendency of a time zone.

  When the utterance extracted by the extraction unit 13 includes a broadcast program name, the second trend analysis unit 15 identifies the utterance regarding the broadcast program with the broadcast program name, and broadcasts the utterance extracted by the extraction unit 13. When the program name is not included, it is specified as a utterance related to the broadcast program broadcast in the utterance time zone of the utterance, the broadcast program corresponding to the specified utterance, and each broadcast program obtained based on the specified utterance Based on the number of utterances, the tendency of broadcast programs viewed by the user is analyzed. Thereby, the information presentation apparatus 10 can obtain the viewing tendency of the broadcast program, particularly the viewing program.

  The registration unit 11 registers SNS account data used by the user. Thereby, the information presentation apparatus 10 can collect the remarks of the user who desires to present information.

  In the above-described embodiment, the information presentation device 10 has been described. However, the present invention is not limited to the information presentation device. That is, the present invention may be a program executed on a computer. Here, a specific example of the computer is the information presentation apparatus 10.

The program executes a collection step, an extraction step, a first trend analysis step, a second trend step, a classification step, and a presentation step.
The collecting step collects comments in the social networking service of users who have an account with the social networking service.
In the extraction step, a message including a broadcast program name or a broadcast station name is extracted from the messages collected in the collecting step.
In the first trend analysis step, the trend of the time zone in which the user views the broadcast program is analyzed based on the comments extracted in the extraction step.
In the second trend analysis step, the trend of the broadcast program viewed by the user is analyzed based on the comments extracted in the extraction step.
The classification step classifies the users according to the viewing tendency based on the analysis result of the first trend analysis step and the analysis result of the second trend analysis step.
The presenting step presents information to the user in accordance with the classification result obtained in the classification step.

  Note that the program may be recorded on a computer-readable recording medium. The “recording medium” refers to a storage device such as a portable medium such as a flexible disk, a magneto-optical disk, a ROM, and a CD-ROM, and a hard disk built in a computer system.

DESCRIPTION OF SYMBOLS 10 Information presentation apparatus 11 Registration part 12 Collection part 13 Extraction part 14 1st tendency analysis part 15 2nd tendency analysis part 16 Classification part 17 Presentation part

Claims (5)

A collection unit that collects comments in a social networking service of a user who has an account in the social networking service;
An extraction unit for extracting a utterance about a broadcast program or a broadcast station from the utterances collected by the collection unit;
A first trend analysis unit that analyzes a trend of a time zone during which a user views a broadcast program based on the remarks extracted by the extraction unit;
A second trend analysis unit that analyzes a trend of a broadcast program viewed by the user based on the remarks extracted by the extraction unit;
Based on the analysis result of the first trend analysis unit and the analysis result of the second trend analysis unit, a classification unit that classifies users according to viewing tendency;
A presentation unit that presents information to the user according to the classification result obtained by the classification unit;
An information presentation device comprising: The first trend analysis unit obtains the speech time zone of the speech and the number of speeches corresponding to the speech time zone from the speech extracted by the extraction unit, so that the user can watch the broadcast program The information presenting apparatus according to claim 1, wherein the tendency of the belt is analyzed. The second trend analysis unit
When the broadcast program name is included in the utterance extracted by the extraction unit, specify the utterance regarding the broadcast program of the broadcast program name,
When the broadcast program name is not included in the utterance extracted by the extraction unit, the utterance is specified as a utterance related to the broadcast program broadcast in the utterance time zone of the utterance,
The trend of the broadcast program which a user views is analyzed based on the broadcast program corresponding to the specified message and the number of messages for each broadcast program obtained based on the specified message. Information presentation device. It has a registration unit that registers data of social networking service accounts used by users,
The information collection device according to any one of claims 1 to 3, wherein the collection unit collects comments in a social networking service of a user who has an account registered by the registration unit. A program executed on a computer connected to a communication network,
A collecting step of collecting remarks in a social networking service of a user who has an account with the social networking service;
An extraction step of extracting a utterance including a broadcast program name or a broadcast station name from the utterances collected by the collecting step;
A first trend analysis step of analyzing a trend of a time zone in which a user views a broadcast program based on the remarks extracted by the extraction step;
A second trend analysis step of analyzing a trend of a broadcast program viewed by the user based on the remarks extracted by the extraction step;
A classification step for classifying users according to viewing trends based on the analysis result of the first trend analysis step and the analysis result of the second trend analysis step;
A presentation step for presenting information to the user in accordance with the classification result obtained by the classification step;
A program that executes.

Images