JP2020149456A - Voting tabulation device and program - Google Patents

Abstract

To reduce a load applied to both of a side executing voting and a side voting for someone.SOLUTION: A voting tabulation device 1 comprises an acquisition part 11 for sending a contribution content including information of a sentence to be disclosed to other users other than a user from a user's terminal and for acquiring a contribution content of each of a plurality of users who use a service for distributing the contribution content to terminals of other users who are registered in advance for browsing the contribution content. An extraction part 13 of the voting tabulation device extracts the contribution content including information indicating that the contribution content is voting, out of the contribution contents acquired by the acquisition part 11. A voting destination estimation part 14 calculates an association degree for quantitatively indicating relationship between each extracted contribution content and information related to each voting object which is an object of voting by the contribution content, and then estimates each voting object of the contribution destination on the basis of the association degree calculated for each voting object. A tabulation part 17 tabulates the number of the votes polled for every voting object on the basis of the voting destination estimated for each of the plurality of contribution contents.SELECTED DRAWING: Figure 1

Description

The present invention relates to a voting counting device and a program.

In order to create a so-called digest version such as a omnibus of a TV program, a viewer may vote to select a popular scene from the programs of the past broadcast times. Posting forms are used for this vote. However, every time a vote is held, it is necessary to prepare a dedicated posting form with a pull-down menu for selecting the voting target, which requires labor on the creator side. Viewers also need to access the dedicated posting form using a web browser and enter their name and email address, which takes time to vote.

Therefore, voting using SNS (Social Networking Service) that does not require them is being considered. The creator can vote without preparing a voting form by simply notifying "Please attach a specific hashtag and post on SNS". In addition, viewers can easily vote from a smartphone or the like using an SNS application (app) such as Twitter (registered trademark). In this way, voting using SNS is considered to be beneficial to both program producers and viewers. However, since voting using SNS is described in free text, the program to vote for may not be clear from the text.

On the other hand, there is a technique for automatically classifying information on social media (see, for example, Patent Documents 1 and 2). Both of these technologies analyze social media posts and categorize them according to purpose. In addition, there is a technique for searching content based on search text (see, for example, Patent Documents 3 and 4) and a technique for identifying a topical program based on writing to SNS (see Patent Document 5). ..

Japanese Unexamined Patent Publication No. 2018-63463 JP-A-2017-201437 JP-A-2017-134675 Japanese Patent No. 6151554 Japanese Unexamined Patent Publication No. 2017-12304

All of the techniques of Patent Documents 1 and 2 described above use machine learning in performing the final classification. Therefore, when using these techniques, the creator side needs to prepare learning data suitable for voting. For example, learning data must be prepared each time a one-time vote with a specific purpose, such as a vote for a special program, is carried out. Further, the techniques of Patent Documents 3 to 5 do not specify the voting destination and count the number of votes.

The present invention has been made in consideration of such circumstances, and provides a voting counting device and a program capable of totaling voting results while reducing the load on both the voting side and the voting side. To do.

One aspect of the present invention is a terminal of another user who is registered in advance to transmit the posted content including text information to be disclosed to a user other than the user from the user's terminal and view the posted content. Of the acquisition unit that acquires the posted content of each of the plurality of users who are using the service that distributes the posted content, and the posted content acquired by the acquisition unit, the information indicating that the vote is voted. For each of the extraction unit that extracts the included posted content and the posted content extracted by the extraction unit, the degree of relevance that quantitatively represents the relationship between the information related to each voting target that is the target of voting by the posted content is calculated. Then, based on the voting destination estimation unit that estimates each voting target of the voting destination based on the relevance calculated for each of the voting targets, and the voting destination estimated for each of the plurality of posted contents. It is a voting counting device characterized by including a counting unit that counts the number of votes for each voting target.

One aspect of the present invention is the above-mentioned voting counting device, in which a plurality of votes are excluded from the voting target of counting the number of votes, except for the voting based on the posted content transmitted from the user whose date and time registered in the service is after a predetermined date and time. It is characterized by further including an exclusion unit.

One aspect of the present invention is the above-mentioned voting counting device, in which a plurality of votes are excluded from the voting target of counting the number of votes, excluding the voting based on the posted content first transmitted from the user or transmitted within a predetermined number of times from the beginning. It is characterized by further including an exclusion unit.

One aspect of the present invention is the above-mentioned voting aggregation device, which comprises an organization vote exclusion unit that performs an organization vote exclusion process for each group of the posted contents having the same voting destination estimated by the voting destination estimation unit. Further provided, the acquisition unit acquires information representing the relationship between the users in the service, and the organization vote exclusion process represents the user who transmitted the posted content by a node, and between the users. The graph creation process for creating a graph structure in which the nodes representing the related users are connected based on the information representing the relationship, and the subgroup for each subgroup which is the node group combined in the graph structure. The sufficiency rate calculation process for calculating the sufficiency rate, which is the ratio of the number of relationships constituting the group to the number of the relationships when all the nodes included in the subgroup are related, and the sufficiency rate are The exclusion process for eliminating duplication of votes by the plurality of users represented by the node group included in the subgroup above the threshold, and the number of relationships from the subgroup whose sufficiency rate is below the threshold It is characterized in that it has a repetitive process of removing less than a predetermined number of nodes and performing the graph creation process for the users who are included in the subgroup with a predetermined number or more of the nodes not removed.

One aspect of the present invention is the above-mentioned voting counting device, in which each voting target is a program, and the voting destination estimation unit is a title of each of a plurality of programs for each posted content extracted by the extraction unit. And the first score that quantitatively represents the relationship with the summary sentence and the second score that quantitatively represents the relationship with the subtitles of each of the plurality of the programs are calculated, and the first score and the first score of each of the programs are calculated. (Ii) The program to be voted on is estimated based on the score.

One aspect of the present invention is a program for causing a computer to function as any of the above-mentioned voting counting devices.

According to the present invention, the voting results can be aggregated while reducing the load on both the voting side and the voting side.

It is a functional block diagram which shows the structure of the voting totaling apparatus by one Embodiment of this invention. It is a figure which shows the user information by the same embodiment. It is a figure which shows the program information by the same embodiment. It is a flow chart which shows the processing of the vote totaling apparatus by the same embodiment. It is a figure which shows the algorithm of the organization vote exclusion processing by the same embodiment. It is a flow chart which shows the organization vote exclusion processing by the same embodiment. It is a figure which shows the graph structure by the same embodiment. It is a figure which shows the graph structure by the same embodiment. It is a figure which shows the graph structure by the same embodiment. It is a figure which shows the graph structure by the same embodiment. It is a figure for demonstrating the counting method of the vote by the same embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a functional block diagram showing a configuration of a voting counting device 1 according to an embodiment of the present invention. The voting counting device 1 can communicate with the SNS system 2 and the program information storage device 3. The voting counting device 1 analyzes the contents of the votes cast using the SNS, estimates the program to be voted on, and counts the number of votes for each program.

The SNS system 2 is a service providing system that provides an SNS service to a user registered as a service user. In the SNS service, information to be disclosed to other service users is transmitted from the service user's terminal to the SNS system 2 and registered. Sending information to be disclosed to other service users from the service user's terminal to the SNS system 2 and registering it is described as posting. In addition, the posted information is described as the posted content. The SNS system 2 is a terminal of another service user who is registered in advance that the poster who is the service user who posted the message permits viewing and wishes to view the information published by the poster. Deliver the posted content to. The other service users that the poster permits the posted content to be viewed may be a specific user or all users who use the SNS service.

In the SNS service that can be used in this embodiment, at least text data can be used for the posted content. As such an SNS service, for example, there is Twitter (registered trademark). Registering in the SNS system 2 to deliver the posted content when a specific user posts is described as "following" the user. For example, the user B who wants to distribute the posted content of the user A is a follower who follows the user A. At this time, if the user A is a follower of the user B, the user A and the user B have a bidirectional relationship. When user B is a follower of user A, but user A is not a follower of user B, there is a one-way relationship from user B to user A.

The program information storage device 3 is composed of, for example, one or more database servers. The program information storage device 3 stores the program information of each program. The program information includes information such as the broadcast date and time of the program, the program title, the program outline, and subtitles.

The voting aggregation device 1 includes an acquisition unit 11, a storage unit 12, an extraction unit 13, a voting destination estimation unit 14, a multiple voting exclusion unit 15, an organization vote exclusion unit 16, and an aggregation unit 17. The voting counting device 1 is realized by one or a plurality of computer devices. When the voting counting device 1 is realized by a plurality of computer devices connected by a network, which of these plurality of computer devices is used to realize each functional unit of the voting counting device 1 can be arbitrary. Further, the same functional unit may be realized by a plurality of computer devices. Further, the vote counting device 1 may include a program information storage device 3. Further, the SNS system 2 may include a vote counting device 1.

The acquisition unit 11 acquires SNS information from the SNS system 2. The SNS information includes user information of each user who is a user of the SNS service. The user information includes a user ID which is information for identifying a user, a date and time when service use is registered, a follower, other users who are following, posted information, and the like. The posted information includes the date and time of posting and the content of the posting. The storage unit 12 stores information used for processing of each unit, such as SNS information acquired by the acquisition unit 11. The extraction unit 13 extracts the posted information of the posted content including the information indicating that it is a vote from the posted information included in the user information acquired by the acquisition unit 11. For example, a hash tag is used as information indicating that it is a vote.

The voting destination estimation unit 14 estimates the program voted by the user from the posted content of each posted information extracted by the extraction unit 13. The voting destination estimation unit 14 uses a conventional search technique using a value that quantitatively represents the relationship between documents for this estimation. The voting destination estimation unit 14 has a first score, which is a quantitative value representing the relationship between the document of the posted content and the title and summary sentence of each program, for each posted information extracted by the extraction unit 13. Calculate the second score, which is the degree of relevance that expresses the relationship between the posted content document and the subtitles of each program as a quantitative value. The voting destination estimation unit 14 estimates the voting destination program for each posted information based on the first score and the second score calculated for each program.

The multiple vote exclusion unit 15 excludes a plurality of votes from the same user when there is a limit of "one vote per person". When the multiple vote exclusion unit 15 detects a plurality of post information in which the same user ID is the poster in the post information extracted by the extraction unit 13, the multiple vote exclusion unit 15 selects one of the plurality of post information. Invalidate voting by unselected post information. That is, the multiple voting exclusion unit 15 excludes the voting destination estimated by the voting destination estimation unit 14 for the invalid post information from the counting target of the number of votes.

In addition, the user may newly acquire the user ID of the SNS service and vote in order to make it appear that the vote is not from the same user. Therefore, the multiple vote exclusion unit 15 invalidates the posted information that meets either or both of the following two conditions among the posted information extracted by the extraction unit 13, and votes based on the posted information to total the number of votes. Exclude from the target. The first condition is that the information is posted by a user who newly registered to use the service after the date and time that goes back a predetermined time from the present. The second condition is that the posted information is the first or within a predetermined number of times (two times, three times, etc.) from the beginning after registering the use of the service.

The organization vote exclusion unit 16 eliminates the organization vote. An organization vote is a vote that a certain organization collectively casts on a specific target. Therefore, first, the organization vote exclusion unit 16 sets the users who voted for the same target as one group, and performs a graph creation process for creating a graph structure showing the relationship between the users for each group by referring to the user information. In the graph structure, each user who is the poster of the posted information extracted by the extraction unit 13 is represented as a node, and further, the user is a follower (following) from the node representing each user. The nodes are connected and represented by arrows in the direction leading to the nodes. By this graph creation process, a subgroup consisting of a group of nodes (node group) connected by an arrow and a node group having no connection with other node groups can be obtained. The organization vote exclusion unit 16 performs a sufficiency rate calculation process for calculating the sufficiency rate for each subgroup. The sufficiency rate is the ratio of the number of arrows included in a subgroup to the number of arrows when all the nodes included in the subgroup are connected in both directions. The organization vote exclusion unit 16 determines that votes from users corresponding to nodes included in subgroups whose sufficiency rate is equal to or higher than the threshold value are organization votes, and eliminates duplication of votes for those organization votes and puts them together. The exclusion process is performed with one vote. The organization vote exclusion unit 16 removes the node of the user who is not related to the organization vote for the subgroup whose sufficiency rate is less than the threshold value, and creates a graph for the user corresponding to the node remaining without being removed. Repeat the process Repeat the process. The user's node that is not related to the voting bloc is, for example, the node with the smallest number of arrows. Further, when the number of users included in the subgroup is small, the organization vote exclusion unit 16 determines that the subgroup itself is a user group not related to the organization vote.

The aggregation unit 17 aggregates the number of votes for each program based on the program of the voting destination specified for each of the plurality of posted information. The totaling unit 17 outputs the total number of votes for each program.

FIG. 2 is a diagram showing an example of user information. The user information includes a user ID, a user name, a registration date and time, a follower, a follow target user, the number of posts, and post information. The user ID is identification information that uniquely identifies the user. The user name is the name of the user. The registration date and time is the date and time when the user registered the use of the service. A follower refers to another user who allows the user to publish the posted content. The follow target user indicates another user whose follower is the user. The number of posts is the number of times the user has posted. The posted information includes the date and time of posting and the content of the posting. When the user retweets, the posted information is added with the information that the retweet is made and the information of the user who posted the retweeted posted information. Retweet is posting the content posted by other users that you are following to publish to your followers. The user information may further include information on attributes related to the user. Attributes include, but are not limited to, for example, address, date of birth, gender, and the like.

FIG. 3 is a diagram showing an example of program information. The program information includes information such as a program ID, a broadcast date and time, a program title, a program outline, and subtitles. The program ID is identification information that uniquely identifies the program. The broadcast date and time indicates the date and time when the program was broadcast. The program title is text data indicating the title and subtitle of the program. The program outline is text data such as a synopsis of the program, the contents of the program, and performers. The subtitle is text data indicating the subtitle of the program.

<Operation of voting counting device 1>
FIG. 4 is a flow chart showing the processing of the voting aggregation device 1. When conducting a program vote, the program producer informs the person to vote by SNS with a specific hashtag. The user who uses the SNS posts the posted content including the sentence describing the program to be voted on and the well-known hashtag to the SNS system 2 from the terminal used by the user. When the SNS system 2 receives the posted content from the user's terminal, the SNS system 2 adds the posting date and time and adds it to the user information.

In step S110, the acquisition unit 11 of the vote totaling device 1 reads SNS information from the SNS system 2 and writes it in the storage unit 12 according to the timing of totaling the votes. The read SNS information may not include user information of a user whose disclosure target is only a specific user. In step S120, the extraction unit 13 is set with a hashtag indicating a vote in the posted content and the posting date and time is included in the voting execution period from the user information included in the SNS information stored in the storage unit 12. Extract the posted information. In the following, this extracted post information will be referred to as voting post information. In step S130, the voting destination estimation unit 14 estimates the voting destination program based on the posted information for voting.

Generally, there is a limit to the number of characters in the content posted on SNS. In addition, the posted content is free description. Therefore, the selection of the wording described in the posted content is often different from that used for the program information. Therefore, the voting destination estimation unit 14 uses a search technique for calculating the relationship between documents as a quantitative value, instead of simple keyword matching. In particular, a technique for searching for relationships between documents by extending them to related terms is desirable. For this search technique, for example, the technique of Patent Document 3 can be used. You can also use Word2Vec or Doc2Vec.

In addition, when the posted content uses words that describe the content of the entire program, such as "Kurage Aquarium Time", or "I was impressed by the word △△ that the keeper XX said." In some cases, words and phrases that describe some scenes of the program are used. Therefore, the voting destination estimation unit 14 sets the search target using the posted content to two types, "program title and summary sentence" and "all subtitles of the program" of each program. The voting destination estimation unit 14 weights the scores obtained as the search results of each of the search targets and then adds them to calculate the final score. The multiple voting exclusion unit 15 estimates the program with the highest final score as the voting destination.

When the technique of the program information search device for searching the content related to the search key disclosed in Patent Document 3 is applied, the voting destination estimation unit 14 performs the search as follows. First, the voting destination estimation unit 14 uses the posted content as a search key. The voting destination estimation unit 14 acquires the words included in the posted content as higher-level related words. The voting destination estimation unit 14 reads out the similarity between the related word and the related word for each higher related word from the concept map showing the related words and the similarity between the related words. Further, the voting destination estimation unit 14 repeats a process of reading out the similarity between the other related words and the related other words from the concept map for each read word a predetermined number of times. The voting destination estimation unit 14 is a string of words in which the words read from the concept map are arranged in order from the upper related words to the lower related words for each lower related word which is a word read from the concept map. Get a path.

The voting destination estimation unit 14 determines the similarity between adjacent words in the path from the upper related word to the lower related word, the number of other words related to the word included in the path, and the importance of the words of the upper related word. Based on, the degree of relevance indicating the height of the relevance to the search key is calculated for each subrelated word. The higher the similarity, the smaller the number of other related words, or the higher the importance, the higher the score. In addition, IDF (Inversed Document Frequency) can be used as the importance of the words of the higher related words. Further, the voting destination estimation unit 14 divides the text data to be searched into words, and identifies each word that matches the upper related word or the lower related word among the words obtained by the division. The voting destination estimation unit 14 multiplies the importance and relevance of each of the identified words, and divides the total of the multiplication results by the logarithm of the number of words obtained by dividing the search target to score. To calculate. As the importance of the words used here, okapi-BM25 and the value of CoM disclosed in JP-A-2015-132899 can be used. CoM is a weighting method based on the idea that words in which more related words (words with high similarity) appear in a sentence are more important words in the sentence.

For example, the voting destination estimation unit 14 has a first search engine and a second search engine using the technique of Non-Patent Document 3. In the first search engine, the search target is the program title and the summary sentence, and in the second search engine, the search target is the subtitle of the program. The voting destination estimation unit 14 inputs the posted content of the voting posting information into these first and second search engines as a search key. From the first search engine is output first score s _t, from the second search engine is output second score s _c. Voting end estimation unit 14, by the following equation (1), for each program to be voting destination candidate, reads them program information stored in program information storage device 3, the first score s _t and second score The weighted sum s of s _c is calculated.

s = αs _t + (1-α) s _c ... (1)

The voting destination estimation unit 14 uses the weighted sum s calculated for the program as a score indicating the possibility of voting for the program. In the search targeting subtitles, since many sentences are included in the search target, it is easy to give a relatively high score to programs other than the target intended for posting. Therefore, it is desirable that the value of α is made larger than 0.5 and the weight of the search for the program title and the summary sentence is higher than that for the search for subtitles. The voting destination estimation unit 14 determines that the program having the largest weighted sum (score) s is the voting destination program estimated from the posted content of the voting posted information.

The voting counting device 1 estimates the voting destination of each voting posted information, and then applies a filter according to the voting implementation procedure. Filters include "exclude multiple votes" and "exclude organization votes". First, in step S140, the multiple vote exclusion unit 15 eliminates multiple votes so as to meet the filtering condition of "one vote per person" (step S140). The voting destination estimation unit 14 reads out the user ID of the user who posted each voting posting information from the user information stored in the storage unit 12. When there are a plurality of voting post information having the same user ID, the voting destination estimation unit 14 selects the oldest or newest posting date and time among the voting post information, and the remainder not selected. Voting using the voting information posted in is excluded from the total number of votes. When the filtering condition is "up to 1 vote per person per day", the voting destination estimation unit 14 detects whether or not there is a plurality of voting posting information with the same user ID on the same posting date. The voting destination estimation unit 14 selects the oldest or newest posting time among the plurality of voting posting information with the same posted date and the same user ID, and the remaining voting information that is not selected. Post information is excluded from the total number of votes.

Further, the multiple voting exclusion unit 15 identifies the voting posting information presumed to have been used to vote a plurality of times using different user IDs, and invalidates the voting by the voting posting information. Specifically, the multiple voting exclusion unit 15 reads out the registration date and time of the user who posted the voting posting information estimated to be the voting destination from the user information stored in the storage unit 12. The multiple voting exclusion unit 15 excludes the voting posted information of users whose registration date and time is after a predetermined time from the present date and time from the counting target of the number of votes. For example, the date and time when the voting announcement was first made may be a date and time retroactive from the present by a predetermined time. In addition, the multiple voting exclusion unit 15 collects voting posting information of users whose user information is set to have voted once or within a predetermined number of times (for example, 2 to 3 times) from the total number of votes. except. Alternatively, the multiple voting exclusion unit 15 determines how many times the voting post information for which the voting destination is estimated belongs to the user's post information by referring to the post information included in the user information. The multiple voting exclusion unit 15 excludes voting posting information from the user's first posting or posting within a predetermined number of times from the beginning from the counting target of the number of votes.

In step S150, the organization vote exclusion unit 16 eliminates the organization vote (step S150). The organization vote exclusion unit 16 pays attention to the relationship between the users who voted for the same program among the users who posted the voting information. The organization vote exclusion unit 16 obtains the information of the follow target user registered in the user information as a relationship between the users. In the case of Twitter (registered trademark), information of other users can be embedded in the posted content to be published to followers by a description called @ -mention. Therefore, the organization vote exclusion unit 16 may consider that there is a relationship from user A to user B when the information of user B is described in the posted content of user A using @ -mention. Further, when the user A retweets the posted content of the user C, the user A may be regarded as having a relationship with the user C (a follower of the user C). In this way, the relationship between users can also be obtained from the posted information. In the following, the case where the follow relationship is used as the relationship between users will be described as an example, but the same procedure can be performed by using the relationship between users obtained from @ -mention, retweet, or the like. The process of excluding the voting bloc will be described later with reference to FIGS. 5 and 6. The organization vote exclusion unit 16 collects votes from a group of users who voted for the organization vote in the same program into one vote.

In step S160, the aggregation unit 17 aggregates the voting destinations of the valid voting posting information obtained by excluding the voting by the same user and the organization vote from the voting posting information by the above processing, and votes for each program. Get a number. The totaling unit 17 outputs the number of votes for each program finally obtained.

FIG. 5 is a diagram showing an algorithm for the voting bloc exclusion processing in step S150 of FIG. 4, and FIG. 6 is a flow chart showing the voting bloc exclusion processing using the algorithm shown in FIG. The organization vote exclusion process will be described in detail with reference to FIGS. 5 and 6. The organization vote exclusion unit 16 performs the organization vote exclusion process shown in FIGS. 5 and 6 by setting each program that can be voted as a program t.

First, in step S210, the tissue votes exclusion section 16, the user Ut to vote the program t processed user _^Ut. Voting for the program t means that the voting destination estimated for the voting information is the program t. In step S220, the tissue votes exclusion section 16, the processing target user U ^- Create a graph structure representing a relationship between _t. Tissue votes exclusion section 16, based on the graph structure created, processed user U ^- divide _t into subgroups that summarizes the users connected to each other at any distance. When the number of subgroups is N, the nth subgroup (n is an integer of 1 or more and N or less) is defined as the subgroup U ^- _tn . Tissue votes exclusion section 16, the process continues subgroup U ^{- 'Initialize} like to NULL _tn.

FIG. 7 is a diagram showing an example of a graph structure. The 10 nodes represented by circles represent different processing target users U ^- _t . These nodes are referred to as users U1 to U10. Tissue votes exclusion section 16, the processing target user U ^- _t referring to the user information of the user represented by, connecting the user U1~U10 by arrows representing the follow-up relationship. Arrow from the user Ux (x is 1 to 10 integer) user Uy (y ≠ x, y is 1 to 10 integer) to the processing target user U represents user Ux ^- _t represents the user Uy Indicates that the user U ^- _t to be processed is being followed (followered). Arrows in both directions indicate that they are following each other. In the graph structure shown in FIG. 7, the number N of subgroups is 1.

In step S230 of FIG. 6, the organization vote exclusion unit 16 selects one subgroup U ^- _tn . In step S240, the tissue votes exclusion section 16, sub-group ^U were selected _- the fill rate _{s n} relationship _tn, is calculated using the following equation (2) and (3). Note that u and v are processing target users U ^- _t included in the subgroup U ^- _tn .

The denominator on the right side of the equation (3) is the number of relationships when there is a bidirectional relationship between all the processing target users U ^- _t included in the subgroup U ^- _tn . If the number of users U ^- _{t to be} processed is _M , the denominator can be calculated by MC ₂ . The numerator on the right side of equation (3) is the number of actual relationships in the subgroup U ^- _tn . For example, if the subgroups shown in FIG. 7, the processing target user U ^- the number M of _t is 10, the number of relationships is 23, is filling ratio _{s n = 23/10 C 2} = 0.256 relation ..

In step S250, the tissue votes exclusion section 16 determines whether the fill rate _{s n} has exceeded the threshold th _s. When it is determined that the threshold value is exceeded (step S250: YES), the organization vote exclusion unit 16 performs the process of step S260. That is, the organization vote exclusion unit 16 considers the selected subgroup U ^- _tn as the organization vote group. Tissue votes exclusion section 16, the processing target user U included in the tissue form the group ^- _t and one vote collectively vote for the program t from each.

On the other hand, when the organization vote exclusion unit 16 determines that the sufficiency ratio sn is _equal to or less than the threshold value th _s (step S250: NO), the organization vote exclusion unit 16 performs the process of step S270. That is, the tissue votes exclusion section 16, the processing target user U ^- the number d _u relationships _t, is calculated by the following equation (4). Bidirectional relationships represented by double-headed arrows in the graph structure are counted assuming that the number of relationships is 2.

In step S280, the tissue votes exclusion section 16, sub-group ^U _- number _{d u} of most interest from _tn is less processed user U ^- removing _t. In step S280, the organization vote exclusion unit 16 determines that the processing target user U ^- _t removed in step S270 does not belong to the organization vote group, and determines that the vote of the processing target user U ^- _t is treated as one vote. ..

In step S290, the tissue votes exclusion section 16, the number _{d u} is less processed user U of most interest ^- subgroup after the removal of the _t ^U _- processed user U included in _tn ^- number of _t threshold determines whether or not more than th _u. Tissue votes exclusion section 16, when judging that does not exceed the threshold th _u (step S290: NO), the process of step S300. That is, the organization vote exclusion unit 16 determines that the processing target user U ^- _t included in the subgroup U ^- _tn does not belong to the organization vote group. That is, the organization vote exclusion unit 16 determines that the votes of the processing target users U ^- _t are not the organization votes, but the votes of each processing target user U ^- _t are treated as one vote.

On the other hand, tissue votes exclusion section 16, sub-group ^U _- processed user included in _tn U ^- if the number of _t is determined to exceed the threshold th _u (step S290: YES), performs the process of step S310 .. That is, the tissue votes exclusion section 16, the process continues subgroup U ^- Add _{tn -} ^'current subgroup ^U to _tn.

After the processing of step S260, step S300 or step S310, the organization vote exclusion unit 16 performs the processing of step S320 and determines whether or not all the subgroups U ^- _ton have been selected. When the organization vote exclusion unit 16 determines that there is an unselected subgroup U ^- _tn (step S320: NO), the organization returns to the process of step S230 and selects the next subgroup U ^- _tn .

When it is determined that all N subgroups U ^- _tn have been selected (step S320: YES), the organization vote exclusion unit 16 performs the process of step S330 and determines whether or not the process is completed. For example, tissue votes exclusion section 16, the process continues subgroup U ^{- _'t1} ~U ^-' either _tN is determined not to be terminated when no is NULL, determining the end when all is NULL. When the organization vote exclusion unit 16 determines that the process is not complete (step S330: YES), the organization vote exclusion unit 16 performs the process of step S340. That is, the tissue votes exclusion section 16, the process continues subgroup U ^{- _'t1} ~U ^-' processes the user contained in _tN target user U ^- and _t, repeats the process from step S220. Then, when it is determined that the organization vote exclusion unit 16 is finished (step S330: NO), the organization vote exclusion unit 16 ends the organization vote exclusion process for the program t.

A specific example of the above-mentioned organization vote exclusion process is shown below. A case where the threshold value th _{s of the} sufficiency rate is 0.5 and the threshold value th _{u of the} number of users is 3 will be described as an example.
For example, when the graph structure shown in FIG. 7 is obtained in step S220, the organization vote exclusion unit 16 has a threshold value th _s (= 0.56) of the sufficiency ratio s _n (= 0.256) calculated in step S240. ) It is determined that it is less than or equal to (step S250: NO). The organization vote exclusion unit 16 counts the number of relationships between users U1 to U10.

FIG. 8 is a diagram showing a graph structure describing the results of counting the number of relationships of each of the users U1 to U10. The numbers described in the nodes of user Ux (x = 1-10) represent the number of user Ux relationships. For example, the user U1 is a number of 4 _{d u} relationship, the number _{d u} relationship of the user U2 is 6.

Tissue votes exclusion section 16, sub-group ^U _- number _{d u} of most interest from _tn is small to remove the _d u = 3 users U6 and U7 (step S270). The organization vote exclusion unit 16 considers the users represented by the users U6 and U7 not to belong to the organization vote group (step S280). The number of users in the subgroup U ^- _tn after removing the users U6 and U7 is 8, which exceeds the threshold number of users _{t u} (= 3) (step S290: YES). Therefore, tissue votes exclusion section 16, remaining user U1 - U5, processed user U the U8～U10 ^- to create a re-graph structure as _t (step S310, the step S320: YES, step S330: NO, step S340, Step S220).

FIG. 9 is a diagram showing a graph structure generated after removing users U6 and U7 having the smallest number of relationships in FIG. The number N of the graph structure subgroups shown in FIG. 9 is 2, and has subgroups U ⁻ _t1 composed of users U1 to U5 and U ⁻ _t2 composed of users U8 to U10. Further, FIG. 9 shows the number of relationships of each user U1 to U5 and U8 to U10.

The sufficiency ratio s _n of the first subgroup U ⁻ _t1 is 0.55 and does not exceed the threshold value th _s (step S240, step S250: NO). Thus, tissue votes exclusion section 16, sub-group ^U _- number _{d u} of most interest from _t1 is small to remove the user U5 (step S270). Subgroup ^U after removal _- number of users _t1 is 4 exceeds the threshold _th u = 3 (step S290: YES). Tissue votes exclusion section 16, sub-group ^U after removal _- processing continuation subgroups user U1~U4 belonging to _t1 U ^{- 'Add} to _t1 (step S310).

The sufficiency ratio s _n of the second subgroup U ⁻ _t2 is 1.00, which exceeds the threshold value th _s (step S240, step S250: YES). Tissue votes exclusion section 16, sub-group ^U _{- t2} a determines that the tissue form a group (step S260). This voting bloc group is referred to as G1. Tissue votes exclusion section 16, the process continues subgroup U ^{- _'t1} to set the user U1~U4 are processed user U ^- again to create a graph structure as _t (step S320: YES, step S330: NO, step S340, step S220).

FIG. 10 is a diagram showing a graph structure created after the graph structure shown in FIG. Note that FIG. 10 also shows the already detected voting bloc group G1. The number N of the subgroups of the graph structure shown in FIG. 10 is 1. Further, FIG. 10 describes the number of relationships of the users U1 to U4 constituting the subgroup U ^- _t1 . The sufficiency ratio s _n of the subgroup U ⁻ _t1 is 0.75, which exceeds the threshold value th _s (step S240, step S250: YES). Therefore, the organization vote exclusion unit 16 determines the subgroup U ^- _t1 as the organization vote group (step S260). This organization vote group is referred to as G2. Tissue votes exclusion section 16, the selection of all subgroups ended (step S320: YES), and processing continues subgroup U ^{- _'t1} because not set subgroups (step S330: YES), End the multiple vote exclusion process. Based on the above, it is determined whether or not each user U1 to U10 belongs to the organization vote group.

FIG. 11 is a diagram for explaining a method of counting votes from users who have obtained the graph structure shown in FIG. 7. The organization vote exclusion unit 16 collects votes from users U1 to U4 belonging to the organization vote group G2 into one vote, and votes from users U8 to U10 belonging to the organization vote group G1 together into one vote. Therefore, the counting unit 17 totals the number of votes for the program t as 5 votes from the organization vote group G2, the user U5, the user U6, the user U7, and the organization vote group G1.

The threshold value th _s fulfillment ratio, and the threshold th _u number of users can be set arbitrarily. Further, when the user information includes the information of the user's attribute, the extraction unit 13 may perform the above processing on the posted information of the user whose attribute matches the predetermined condition. This makes it possible to filter the users who can vote by attribute. In addition, the counting unit 17 may total the number of votes for each attribute of the user.

In the above-described embodiment, the case where the voting target is a program has been described as an example, but the voting target can be arbitrary. In this case, the voting counting device 1 uses text data describing the content related to the voting target instead of the program information. For example, when the voting target is a product, the voting counting device 1 calculates the score by using the product name and the product introduction instead of the program title and the summary sentence, and using the product review instead of the subtitles.

According to the above-described embodiment, the voting counting device 1 can estimate from the text posted on the SNS which program the viewer was trying to vote for. Therefore, since it is possible to carry out program voting using SNS, the voting side does not need to prepare a posting form, and the viewer can easily vote. In addition, the counting of votes by the voting counting device 1 can also be used for voting that is carried out only once, such as voting for a certain special number.

Further, as a feature of posting using SNS, since the information of the poster can be used, the voting counting device 1 can perform filtering according to the voting recruitment requirements. For example, since viewers can easily vote, there is a possibility that multiple votes from the same user and organization votes may occur. The vote counting device 1 can refer to the information of the poster on the SNS and perform filtering such as limiting "only one vote per person" and excluding the organization vote, which was difficult to post from the posting form. it can.

The functions of the voting counting device 1 and the SNS system 2 in the above-described embodiment may be realized by a computer. In that case, a program for realizing the functions of the vote counting device 1 and the SNS system 2 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read by the computer system and executed. It may be realized by. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, or a storage device such as a hard disk built in a computer system. Further, a "computer-readable recording medium" is a communication line for transmitting a program via a network such as the Internet or a communication line such as a telephone line, and dynamically holds the program for a short period of time. It may also include a program that holds a program for a certain period of time, such as a volatile memory inside a computer system that serves as a server or a client in that case. Further, the above-mentioned program may be a program for realizing a part of the above-mentioned functions, and may be a program for realizing the above-mentioned functions in combination with a program already recorded in the computer system.

Although the embodiments of the present invention have been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and the design and the like within a range not deviating from the gist of the present invention are also included.

1 ... Voting totaling device, 2 ... SNS system, 3 ... Program information storage device, 11 ... Acquisition unit, 12 ... Storage unit, 13 ... Extraction unit, 14 ... Voting destination estimation unit, 15 ... Multiple voting exclusion unit, 16 ... Organization Voting exclusion department, 17 ... Aggregation department

Claims (6)

The posted content including the information of the text to be disclosed to other users other than the user is transmitted from the user's terminal, and the posted content is delivered to the terminal of another user who is registered in advance to view the posted content. An acquisition unit that acquires the posted content of each of the plurality of users who are using the service
From the posted contents acquired by the acquisition unit, an extraction unit that extracts the posted contents including information indicating that it is a vote, and an extraction unit.
For each of the posted contents extracted by the extraction unit, the degree of relevance that quantitatively represents the relationship with the information about each voting target to be voted by the posted contents is calculated, and the calculated degree is calculated for each of the voting targets. A voting destination estimation unit that estimates each of the voting targets of the voting destination based on the degree of relevance,
A totaling unit that aggregates the number of votes for each voting target based on the voting destination estimated for each of the plurality of posted contents.
A voting counting device characterized by being equipped with. It further includes a multiple voting exclusion unit that excludes votes based on the posted content sent from the user whose date and time registered in the service is after a predetermined date and time from the target of counting the number of votes.
The voting counting device according to claim 1, wherein the voting counting device is characterized. It further includes a multiple voting exclusion unit that excludes votes based on the posted content first transmitted from the user or transmitted within a predetermined number of times from the beginning from the target of counting the number of votes.
The voting counting device according to claim 1, wherein the voting counting device is characterized. Further provided with an organization vote exclusion unit that performs organization vote exclusion processing for each group of the posted contents having the same voting destination estimated by the voting destination estimation unit.
The acquisition unit acquires information representing the relationship between the users in the service, and obtains information.
The organization vote exclusion process is
A graph creation process for creating a graph structure in which the users who have transmitted the posted content are represented by nodes, and the nodes representing the related users are connected based on the information representing the relationship between the users. ,
For each subgroup that is the node group combined in the graph structure, the number of relationships that make up the subgroup is the number of relationships when all of the nodes included in the subgroup are related. Satisfaction rate calculation processing to calculate the sufficiency rate, which is the ratio of occupancy,
Exclusion processing that eliminates duplication of voting by the plurality of users represented by the node group included in the subgroup whose sufficiency rate is equal to or higher than the threshold value.
The target is the user who removes nodes having a smaller number of relationships than a predetermined number from the subgroup whose sufficiency rate is equal to or lower than the threshold value, and includes the unremoved nodes in the subgroup having a predetermined number or more. Has an iterative process that performs graph creation processing,
The voting counting device according to any one of claims 1 to 3, wherein the voting counting device is characterized by the above. Each of the voting targets is a program
The voting destination estimation unit includes a first score that quantitatively represents the relationship with the titles and summary sentences of each of the plurality of programs for each of the posted contents extracted by the extraction unit, and subtitles of each of the plurality of programs. A second score that quantitatively represents the relationship is calculated, and the program to be voted on is estimated based on the first score and the second score of each of the programs.
The voting counting device according to any one of claims 1 to 4, wherein the voting counting device is characterized by the above. A program for causing a computer to function as the voting counting device according to any one of claims 1 to 5.

Images