JP2017084225A - User type determination device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To identify a local user who transmits information on a particular area.SOLUTION: A user type determination device 20 comprises: a reception unit 22 for obtaining SNS information including profile information on the profile of a user and document information on a document contributed by the user; a geographical locality calculation unit 24 for calculating a score Sl on the geographical locality of the user on the basis of the SNS information; and a determination unit 27 for determining whether the user is a local user on the basis of the score Sl.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a user type determination device.

  2. Description of the Related Art Conventionally, there is a user classification device that classifies users based on information acquired from an SNS server in SNS (Social Networking Service) such as Twitter (see, for example, Patent Document 1). The user classification device described in Patent Literature 1 is a user who mainly posts articles, a user who mainly browses articles posted by other users, and a user who mainly re-posts articles posted by other users In addition, each user is classified into a user who mainly posts articles and reposts articles posted by other users.

JP 2012-212286 A

  In SNS, not only general users but also various users such as celebrities, news distribution, sightseeing spots, and commercial facilities post highly immediate information. For example, a user (account) related to facilities such as tourist spots and commercial facilities may post event information, sale information, and the like. A local user who transmits information about a specific region, such as a user about a facility, may post useful information for other users. However, the user classification device described in Patent Literature 1 cannot identify a regional user.

  The present invention provides a user type determination apparatus capable of specifying a regional user who transmits information related to a specific region.

  Whether the user type determination device according to one aspect of the present invention is a regional user in which a plurality of users each post a document and a user who uses a communication service for viewing the posted document transmits information on a specific region It is a user type determination apparatus which determines whether or not. The user type determination device includes acquisition means for acquiring service information including profile information related to a user's profile and document information related to a document posted by the user, and a location included in information transmitted by the user based on the service information. Geographic locality calculation means for calculating a first feature quantity related to geographic locality indicating the degree of limitation of the geographical range, and whether or not the user is a regional user is determined based on the first feature quantity. Determining means.

  In this user type determination device, a first feature amount related to a user's geographical locality is calculated based on service information including profile information related to a user's profile and document information related to a document posted by the user. Then, based on the first feature amount, it is determined whether or not the user is a regional user who transmits information related to a specific region. Local users often post documents that describe information within a specific region. For this reason, regional users are likely to have greater geographical locality than other users. Therefore, for example, when the first feature amount is high, the user can be determined to be a regional user, and the regional user can be specified.

  The geographical locality calculation unit may acquire one or more pieces of position information based on the service information, and may calculate the first feature amount based on a distribution of the acquired one or more pieces of position information. Regional users often have a profile for a specific region and often post documents that describe information within a specific region. For this reason, there is a high possibility that the dispersion of the location information included in the profile of the regional user and the posted document is small. Thus, the smaller the variance of the location information acquired from the service information, the greater the geographical locality, and the greater the variance, the smaller the geographic locality. Therefore, the variance of the location information acquired from the service information is It can be an indicator of geographical locality. Therefore, the first feature amount related to the geographical locality of the user can be calculated based on the distribution of one or more pieces of position information acquired from the service information.

  The user type determination device according to another aspect of the present invention may further include a formality calculation unit that calculates a second feature amount related to the formality indicating the reliability of the information transmitted by the user based on the service information. . The determination unit may determine whether the user is a regional user based on the first feature value and the second feature value. When the user is an official user, the reliability of the information included in the document posted by the user is high. For this reason, it is possible to specify a regional user who transmits highly reliable information regarding a specific region by determining whether or not the user is a regional user in consideration of officialness.

  The officialness calculation means calculates the second feature amount based on the number of links indicating the number of other users linked to the user and the number of linked users indicating the number of other users linked to the user. May be. It is considered that a user who transmits useful information, that is, reliable information, is more likely to be linked from other users than the number linked to other users. For this reason, since the number of links and the number of links can be an index of the user's officialness, it is possible to calculate the second feature amount related to the user's officiality based on the number of links and the number of linked links.

  The user type determination device according to still another aspect of the present invention further includes a sociality calculation means for calculating a third feature value related to sociality indicating the degree to which a user communicates with another user based on service information. May be. The determination unit may determine whether the user is a regional user based on the first feature value, the second feature value, and the third feature value. An official user has a high possibility of not having a conversation with other users individually, and a possibility of low sociality is high. For this reason, it is possible to improve the accuracy of determining whether or not the user is an official user by considering sociality.

  The sociality calculation means may calculate the third feature amount based on the number of documents posted by the user toward other users among the documents posted by the user. Since the number of documents that the user has posted to other users can be regarded as the number of conversations that the user has individually with other users, it can be an index indicating the sociality of the user. For this reason, the 3rd feature-value regarding a user's social property is computable based on the number of documents which a user contributed to other users.

  The determination unit determines that the user is a regional user when the first feature value is equal to or greater than the first threshold value, the second feature value is equal to or greater than the second threshold value, and the third feature value is equal to or less than the third threshold value. May be. In this case, a user with high geographical locality and officialness and low sociality is determined as a regional user. Since the information transmitted by such a regional user is reliable information regarding a specific region, it is possible to identify a regional user who transmits highly reliable information regarding a specific region.

  The user type determination device according to still another aspect of the present invention may further include a regional user information storage unit that stores regional user information related to the regional user. When the determination unit determines that the user is a regional user, the determination unit may add the regional user information to the regional user information storage unit based on the user profile information. By storing information related to regional users, it is possible to easily search for regional users who transmit information related to various regions.

  ADVANTAGE OF THE INVENTION According to this invention, the area user who transmits the information regarding a specific area can be specified.

It is a figure which shows the structure of a user classification determination system. It is a block diagram which shows the function structure of the user classification determination system of FIG. It is a figure which shows the hardware constitutions of the user classification determination apparatus of FIG. It is a figure which shows an example of a SNS information table. It is a figure for demonstrating the characteristic for every user classification. It is a figure which shows an example of a local user information table. It is a figure which shows an example of the service by the user classification determination apparatus of FIG. It is a flowchart which shows a series of processes of the user type determination method which the user type determination apparatus of FIG. 1 performs. It is a flowchart which shows the detail of the SNS information acquisition process of FIG. It is a flowchart which shows the detail of the geographical locality calculation process of FIG. It is a flowchart which shows the detail of the officialness calculation process of FIG. It is a flowchart which shows the detail of the social calculation process of FIG. It is a flowchart which shows the detail of the user determination process of FIG.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same elements are denoted by the same reference numerals, and redundant description is omitted.

  FIG. 1 is a diagram illustrating a configuration of a user type determination system. As shown in FIG. 1, the user type determination system 1 is a system for specifying a regional user among users (accounts) using a service provided by the SNS server 10. Device 20. Here, the service provided by the SNS server 10 is a so-called SNS (Social Networking Service) in which a plurality of users each post a document and browse the posted document. In this embodiment, a description will be given using Twitter as the SNS.

  A regional user is a user who transmits information about a specific region. A region means a region within a certain range, for example, a municipality and a section obtained by dividing a region on a map. The account is an authority for using a service provided by the SNS server 10, and may be used to mean identification information of a user having the authority. That is, the user who uses SNS has an account.

  The SNS server 10 is a server device that stores SNS information (service information) including user profile information and document information. Profile information is information relating to the user's profile. The document information is information related to the document posted by the user. Details of the profile information and document information will be described later. The user type determination device 20 is a device that determines whether or not a target user who is a determination target user is a regional user. The user type determination device 20 and the SNS server 10 are connected via a network N so that they can communicate with each other.

  Next, the functional configuration of the user type determination system 1 will be described with reference to FIG. FIG. 2 is a block diagram illustrating a functional configuration of the user type determination system 1. As shown in FIG. 2, the SNS server 10 functionally includes a receiving unit 11, a response unit 12, a profile information DB 13, and a user posting DB 14.

  The receiving unit 11 functions as a receiving unit that receives the SNS information acquisition request transmitted by the user type determination device 20. The SNS information acquisition request is a request for acquiring SNS information stored in the SNS server 10. When receiving the SNS information acquisition request, the reception unit 11 outputs the SNS information acquisition request to the response unit 12.

  The response unit 12 functions as a response unit that responds to the SNS information acquisition request. When the response unit 12 receives the SNS information acquisition request from the reception unit 11, the response unit 12 acquires profile information from the profile information DB 13 and acquires document information from the user posting DB 14. The response unit 12 transmits the acquired profile information and document information to the user type determination device 20 as SNS information.

  The profile information DB 13 functions as profile information storage means for storing profile information of each user. Profile information is stored for each user. For example, user ID (Identification), user name, screen name, number of followers, number of followers, total number of posts, description information, location information, URL (Uniform Resource Locator), authentication Information (Verified) and profile images are included.

  The user ID is identification information that can uniquely identify the user. The user name is a character string indicating the name of the user. The screen name is a character string indicating a name different from the user name, and is expressed in a format of “@xxxx” on Twitter. The number of follow is the number of other users to whom the user is linked (the number of links), and is the number of other users that the user is following on Twitter. The number of followers is the number of other users linking users (the number of linked users), and is the number of other users following the user on Twitter. The total number of posts is the number of times the user has posted a document.

  The description information is a character string indicating a self-introduction sentence, and is arbitrarily set by the user. The location information is a character string indicating a place such as an address, and is arbitrarily set by the user. The URL is, for example, a character string indicating the URL of a Web page that the user has, and is arbitrarily set by the user. The authentication information is information indicating whether or not the user is an authenticated user. When the authentication information is “True”, the user is authenticated, and when the authentication information is “False”, the user is not authenticated. The profile image is an image for identifying the user and is arbitrarily set by the user. The profile information DB 13 may store the URL of the profile image as a profile image.

  The user posting DB 14 functions as a document information storage unit that stores user document information. This document is also called a tweet or a comment depending on the SNS. The document information is stored for each posted document, and includes, for example, a status ID, a user ID, a posting date, a posted document, and a geotag. The status ID is identification information that can uniquely identify a posted document. The posting date / time is information indicating the date / time when the document was posted. The posted document is a document posted by the user. The geotag is position information added to the posted document, and indicates, for example, the latitude and longitude of the place where the document is posted.

  Next, functions of the user type determination device 20 will be described. As shown in FIG. 2, the user type determination device 20 functionally includes a request unit 21, a reception unit 22, an SNS information DB 23, a geographical locality calculation unit 24, and a formality calculation unit 25. The social calculation unit 26, the determination unit 27, and the local user information DB 28 are provided. The user type determination device 20 is configured by, for example, hardware shown in FIG.

  FIG. 3 is a diagram illustrating a hardware configuration of the user type determination device 20. As shown in FIG. 3, the user type determination device 20 physically includes one or more CPUs (Central Processing Units) 201, a RAM (Random Access Memory) 202 as a main storage device, and a ROM (Read Only Memory). 203, a computer including a communication module 204, which is a data transmission / reception device, and an auxiliary storage device 205 such as a hard disk and a flash memory. Each function of the user type determination device 20 shown in FIG. 2 operates the communication module 204 under the control of the CPU 201 by reading one or a plurality of predetermined computer software on the hardware such as the CPU 201 and the RAM 202. In addition, it is realized by reading and writing data in the RAM 202 and the auxiliary storage device 205. The SNS server 10 shown in FIG. 2 is also configured as a computer system similar to the user type determination device 20.

  With reference to FIG. 2 again, details of each function of the user type determination device 20 will be described. The request unit 21 functions as a request unit that requests the SNS server 10 for SNS information. Specifically, the request unit 21 transmits an SNS information acquisition request to the SNS server 10. The SNS information acquisition request is made using, for example, an API (Application Programming Interface).

  The receiving unit 22 functions as an acquisition unit that acquires SNS information. The receiving unit 22 receives the SNS information transmitted by the SNS server 10 and outputs the received SNS information to the SNS information DB 23 for storage. When the URL of the image file is used as the profile image included in the profile information, the receiving unit 22 may acquire the profile image by accessing the URL.

  The SNS information DB 23 functions as an SNS information storage unit that stores SNS information acquired from the SNS server 10. The SNS information DB 23 has an SNS information table that stores SNS information. FIG. 4 is a diagram illustrating an example of the SNS information table. As shown in FIG. 4, the SNS information table includes a status ID, a user ID, a user name, a screen name, a posting date, a posted document, a geotag, a follow number, a follower number, and a total number. SNS information in which the number of posts, description information, location information, URL, authentication information, profile image URL, and profile image are associated with each other is included.

  SNS information is stored for each posted document. That is, when the same user posts a plurality of documents, the SNS information for each document is stored in the SNS information table. The description information, the location information, the URL, the profile image URL, and the profile image are information arbitrarily set by the user. Therefore, if the user has not set these information, the description information, the location information, the URL, and the profile image are included in the SNS information. Absent.

  The geographic locality calculation unit 24 functions as a geographic locality calculation unit that calculates a score S1 (first feature amount) related to the geographic locality of the target user based on the SNS information. The geographical locality means the degree of limitation of the geographical range including the location included in the information (document information and profile information) transmitted by the user. The higher the geographical locality, the more geographically limited. Yes. A larger score S1 indicates a higher geographical locality. A method for calculating the score Sl will be described below.

  For example, the geographical locality calculation unit 24 acquires one or more pieces of position information based on the SNS information stored in the SNS information DB 23. The geographical locality calculation unit 24 has a location information table in which location information indicating a location such as a facility name and a place name is associated with location information indicating the location (for example, latitude and longitude) of the location. Yes. The location information table is set in advance. The geographic locality calculation unit 24 performs, for example, the following analysis and acquires position information regarding the target user.

  The geographical locality calculation unit 24 acquires the description information and the location information associated with the user ID of the target user from the SNS information DB 23. The geographical locality calculation unit 24 acquires, for example, the description information and the location information associated with the latest posting date and time among the description information and the location information associated with the user ID of the target user. The geographical locality calculation unit 24 performs morphological analysis on the description information and the location information, and extracts a character string that can specify a location such as a facility name, a place name, and an address from the description information and the location information. The geographic locality calculation unit 24 refers to the location information table and acquires position information associated with the location specified by the extracted character string.

  The geographical locality calculation unit 24 acquires a posted document associated with the user ID of the target user from the SNS information DB 23. The geographical locality calculation unit 24 performs a morphological analysis on the posted document, and extracts a character string that can specify a location such as a facility name, a place name, and an address from the posted document. The geographic locality calculation unit 24 refers to the location information table and acquires position information associated with the location specified by the extracted character string.

  The geographical locality calculation unit 24 acquires the geotag of the posted document associated with the user ID of the target user from the SNS information DB 23 and uses the geotag as position information.

  The geographical locality calculation unit 24 calculates the score Sl based on the position information acquired from the SNS information. Regional users often have a profile for a specific region and often post documents that describe information within a specific region. For example, a facility official user who is an official user of a facility such as a tourist spot or a commercial facility may include a location including the location of the facility in the location information or a document including information on events in and around the facility. is there. For this reason, there is a high possibility that the dispersion of the location information included in the profile of the regional user and the posted document is small. As described above, the smaller the variance of the position information acquired from the information included in the SNS information, the greater the geographical locality, and the greater the variance, the smaller the geographical locality. Therefore, one or more positions acquired from the SNS information. The distribution of information can be an indicator of geographic locality.

For this reason, the geographical locality calculation unit 24 scores based on the variance of N (N is an integer of 1 or more) pieces of position information x _i (i is an integer of 1 to N) acquired as described above. Calculate Sl. Specifically, the geographic locality calculation unit 24 calculates the reciprocal of the variance (unbiased variance) of the N pieces of position information x _i as shown in the equation (1), and the calculation result is set as the score Sl. To do. The average value x _m is an average value of N pieces of position information x _i . If the position information xl from location information is acquired may be position information xl is used as an average value _{x m.} The geographical locality calculation unit 24 outputs the calculated score Sl together with the user ID of the target user to the determination unit 27.

  When the location information xl is acquired from the location information, the geographic locality calculation unit 24 outputs the location information xl together with the user ID of the target user to the determination unit 27.

  The formality calculation unit 25 functions as a formality calculation unit that calculates a score So (second feature amount) related to the officialness of the target user based on the SNS information. The formality means the reliability of information transmitted by the user, and the higher the formality, the higher the reliability. The higher the score So, the higher the officialness. A method for calculating the score So will be described below.

For example, the officialness calculation unit 25 acquires the number of followers, the number of followers, the description information, and the profile image associated with the user ID of the target user from the SNS information DB 23. The officialness calculation unit 25 sets the follow number associated with the latest posting date and time among the acquired follow numbers as the follow number C_follow, and sets the number of followers associated with the latest posted date and time among the acquired followers. The number of followers is C_follower. The formula calculation unit 25 calculates the ratio Rf of the number of followers to the number of followers, as shown in Expression (2). Users who send useful information, that is, reliable information, are more likely to be followed by other users than they follow other users. For this reason, the follow number C_follow, the follower number C_follower, and the ratio Rf can be indicators of officialness.

  The officialness calculation unit 25 calculates a logo degree Pl indicating the possibility that the profile image is a logo from the profile image. The logo is a figure represented by combining names and marks of companies and facilities in order to impress the image of the companies and facilities. The larger the logo degree Pl, the higher the possibility of being a logo. Official users such as corporate users and facility official users often use corporate logos and facility logos as profile images. For this reason, the logo degree Pl can be an index of officialness.

  The officialness calculation unit 25 calculates a weighted sum of pixel values of each pixel of the profile image using, for example, a neural network, and sets the calculation result as the logo degree Pl. Further, the officialness calculation unit 25 may calculate the logo degree Pl using Bag-of-Visual-words. In this case, the formality calculation unit 25 calculates SIFT (Scale-Invariant Feature Transform) feature amounts of pixels corresponding to pixels or feature points at regular intervals in the profile image, and clusters the SIFT feature amounts. Let the number of SIFT feature values be feature values. The officialness calculation unit 25 calculates the logo degree Pl from the feature amount using SVM (Support Vector Machine) or the like.

  The logo is characterized in that the edge in the profile image is clear, the number of straight lines included in the profile image is large, and the number of colors used in the profile image is small compared to a photograph or the like. In the calculation of the logo degree Pl, the logo degree Pl is calculated so that the logo degree Pl becomes large in the case of having these characteristics. In addition, the officialness calculation part 25 may calculate logo degree Pl beforehand about the profile image memorize | stored in SNS information DB23 irrespective of a target user's determination process.

  The officialness calculation unit 25 extracts, from the acquired description information, a character string that indicates the facility like business hours and an inquiry destination using a regular expression. Examples of the inquiry destination include a telephone number and an e-mail address. For example, “business hours”, “** hours to ** hours”, “**: ** to **: **”, and the like are used as the character strings of the business hours. As the character string of the telephone number, for example, “telephone number”, “0 *-****-****”, “0 ***-***-***”, and the like are used. For example, “E-mail” and “@ ***. Co.jp” are used as the character string of the mail address. “*” Indicates an arbitrary character.

  The officialness calculation unit 25 calculates the facility degree Ps based on the presence or absence of information indicating the facility-likeness. For example, the officialness calculation unit 25 calculates the facility level Ps by adding “1” when character strings corresponding to business hours, telephone numbers, and e-mail addresses are extracted. In this case, the facility degree Ps takes a value of 0 to 3.

The officialness calculation unit 25 calculates the score So based on the follow number C_follow, the follower number C_follower, the ratio Rf, the logo degree Pl, and the facility degree Ps calculated as described above. Specifically, the officialness calculation unit 25 calculates the weighted sum of each value as shown in Expression (3), and sets the calculation result as the score So. The weights w1 to w5 are predetermined values. The greater the influence each element has on the determination of officialness, the greater the weight assigned to that element. The officialness calculation unit 25 outputs the calculated score So to the determination unit 27 together with the user ID of the target user.

  The sociality calculation unit 26 functions as a sociality calculation unit that calculates a score Ss (third feature amount) related to the sociality of the target user based on the SNS information. Sociality means the degree to which a user communicates with other users, and the higher the sociality, the higher the frequency of communication with other users. A larger score Ss indicates higher sociality. A method for calculating the score Ss will be described below.

  The sociality calculation unit 26 acquires, for example, the posted document and the total number of posts associated with the user ID of the target user from the SNS information DB 23. The sociality calculation unit 26 sets the total number of posts associated with the latest posting date and time among the acquired total number of posts as the total number of posts Ca. Of the documents posted by the target user, the number of documents posted by the target user toward other users can be regarded as the number of individual conversations with the other users, so an index indicating sociality Can be.

The sociality calculation unit 26 calculates the number of conversations Cm of the target user based on the acquired posted document. The number of conversations Cm is the number of documents (that is, conversations) posted by the target user toward other users. In Twitter, by including “@user name” in the document, a conversation with the user with the user name becomes possible. For this reason, for example, the social calculation unit 26 analyzes the posted document, counts the number of posted documents including the at sign “@”, and sets the counted number as the number of conversations Cm. The sociality calculation unit 26 calculates the ratio of the conversation number Cm of the target user with respect to the total posting number Ca, as shown in the equation (4), and sets the calculation result as the relative conversation number Cr.

The sociality calculation unit 26 calculates the score Ss based on the number of documents posted to other users among the documents posted by the target user. Specifically, the sociality calculation unit 26 calculates the weighted sum of each value as shown in the equation (5) using the conversation number Cm and the relative conversation number Cr calculated as described above, and The calculation result is a score Ss. Note that the weight w6 and the weight w7 are predetermined values. The greater the influence that each element has on the determination of sociality, the greater the weight assigned to that element. The sociality calculation unit 26 outputs the calculated score Ss to the determination unit 27 together with the user ID of the target user.

  The determination unit 27 functions as a determination unit that determines whether or not the target user is a regional user. As shown in FIG. 5, geographical locality, officialness, and sociality are different for each type of user. Users include general users, celebrities, facility official users, news distribution users, and the like. For example, the regularity of a general user is lower than that of other users. The facility official user is likely to include the facility location and contact information in the profile information, and is more likely to send information about the facility and the surroundings of the facility and not to have individual conversations with other users . For this reason, the officialness and geographical locality of the facility official user are high, and the sociality of the facility official user is low.

  Since the news distribution user aims to distribute information, there is a high possibility that the news distribution user does not individually talk with other users. Moreover, a news delivery user has a high possibility of delivering news in various regions. For this reason, the news distribution user has high officialness, but the news distribution user has low geographical locality and sociality. Since celebrities may have conversations with other users individually and may be active in various regions except for local celebrities, celebrities are highly official and social, and the celebrity's geographical location The nature is low. Thus, the type of user can be classified according to geographical locality, formality and sociality.

  Therefore, the determination unit 27 determines whether the target user is a regional user based on the score S1. For example, the determination unit 27 determines that the target user is a regional user when the score S1 is equal to or greater than the threshold value Sth1 (first threshold value), and the target user is the regional user when the score S1 is smaller than the threshold value Sth1. It is determined that it is not. The regional users determined in this way include general users who transmit information regarding a specific region. Information transmitted by general users may include information with low reliability.

  For this reason, the determination unit 27 may determine whether the target user is a regional user based on the score So in addition to the score S1. For example, when the score S1 is equal to or greater than the threshold value Sth1 and the score So is equal to or greater than the threshold value Sth2 (second threshold value), the determination unit 27 determines that the target user is a regional user, and the score S1 is equal to the threshold value Sth1. If the score So is smaller than the threshold value Sth2, it is determined that the target user is not a regional user. Users with high geographical locality and officialness include facility official users and local celebrities. Since these users are official users, there is a high possibility of transmitting highly reliable information.

  Among these, the facility official user often transmits information about the facility and the surroundings of the facility, and particularly often transmits information useful for other users such as event information and sale information. Therefore, the determination unit 27 may determine whether or not the target user is a regional user (facility official user) based on the score S1, the score So, and the score Ss. For example, when the score S1 is equal to or greater than the threshold Sth1, the score So is equal to or greater than the threshold Sth2, and the score Ss is equal to or less than the threshold Sth3 (third threshold), the determination unit 27 determines that the target user is a regional user (facility It is determined that the user is an official user. When the score S1 is smaller than the threshold value Sth1, the score So is smaller than the threshold value Sth2, or when the score Ss is larger than the threshold value Sth3, the determination unit 27 determines that the target user is not a regional user (facility official user). judge.

  The determination unit 27 may determine whether the target user is a regional user by machine learning using the score S1, the score So, and the score Ss.

  When the determination unit 27 determines that the target user is a regional user, the determination unit 27 adds information on the target user to the regional user information DB 28 as regional user information based on the profile information of the target user. Specifically, the determination unit 27 acquires a user name and a screen name associated with the user ID of the target user from the SNS information DB 23. Then, the determination unit 27 associates the acquired user name and screen name with the user ID and stores them in the regional user information DB 28 as regional user information. When the determination unit 27 receives the position information xl together with the user ID of the target user from the geographical locality calculation unit 24, the determination unit 27 further adds the position information xl to the local user information associated with the user ID.

  The regional user information DB 28 functions as a regional user information storage unit that stores regional user information related to regional users. The regional user information DB 28 has a regional user information table that stores regional user information. FIG. 6 is a diagram illustrating an example of the regional user information table. As shown in FIG. 6, the regional user information table includes regional user information in which a user ID, a user name, a screen name, and position information are associated with each other. Note that when the location information xl is not acquired from the location information of the target user, the regional user information does not include location information.

  The user type determination device 20 may provide a service for distributing peripheral facility information regarding facilities existing around the mobile terminal to the user of the mobile terminal using the regional user information DB 28. In this case, the user type determination device 20 may further include a request reception unit 31 and a provision unit 32.

  The request receiving unit 31 functions as a request receiving unit that receives a peripheral facility information request for requesting peripheral facility information. For example, when an application for searching for a peripheral facility is installed on the mobile terminal, if the peripheral facility is searched for on the mobile terminal, the peripheral facility information request is sent from the mobile terminal together with the location information of the mobile terminal to the user type. It is transmitted to the determination device 20. When the request receiving unit 31 receives the peripheral facility information request together with the position information, the request receiving unit 31 outputs the position information and the peripheral facility information request to the providing unit 32.

  The providing unit 32 functions as providing means for providing peripheral facility information to the mobile terminal. When receiving the location information and the surrounding facility information request from the request receiving unit 31, the providing unit 32 refers to the regional user information table in the regional user information DB 28, and the local user having the positional information within a predetermined distance from the position of the mobile terminal. Extract regional user information. For example, the regional user determined using the score S1, the score So, and the score Ss is likely to be a facility official user. For this reason, the providing unit 32 transmits the extracted regional user information as peripheral facility information to the portable terminal that transmitted the peripheral facility information request.

  As illustrated in FIG. 7, for example, a user (account) of a peripheral facility is displayed on the mobile terminal based on the information transmitted by the providing unit 32. By displaying the user of the peripheral facility on the mobile terminal, the user of the mobile terminal can know events and the like performed in the vicinity.

  Next, processing of the user type determination method in the user type determination device 20 will be described with reference to FIGS. FIG. 8 is a flowchart showing a series of processes of the user type determination method performed by the user type determination device 20. FIG. 9 is a flowchart showing in detail the SNS information acquisition process of FIG. FIG. 10 is a flowchart showing in detail the geographic locality calculation process of FIG. FIG. 11 is a flowchart showing in detail the officialness calculation process of FIG. FIG. 12 is a flowchart showing the sociality calculation process of FIG. 8 in detail. FIG. 13 is a flowchart showing in detail the user determination process of FIG. The process illustrated in FIG. 8 is started, for example, at a constant cycle (for example, every 5 to 10 minutes).

  First, the user type determination device 20 performs an SNS information acquisition process (step S01). In the SNS information acquisition process in step S01, as shown in FIG. 9, first, the request unit 21 transmits an SNS information acquisition request to the SNS server 10 (step S11). And the receiving part 22 receives SNS information from the SNS server 10 (step S12), outputs the received SNS information to SNS information DB23, and stores it (step S13).

  Furthermore, when the URL of the image file is used as the profile image included in the profile information, the receiving unit 22 acquires the profile image by accessing the URL, and outputs the acquired profile image to the SNS information DB 23. (Step S14). As described above, the SNS information acquisition process ends, and the process proceeds to step S02.

  Subsequently, the user type determination device 20 performs a geographical locality calculation process (step S02). In the geographical locality calculation process in step S02, as shown in FIG. 10, first, the geographical locality calculation unit 24 performs the description information, the location information, the posted document, and the correspondence information associated with the user ID of the target user. A geotag is acquired from SNS information DB23 (step S21).

  Then, the geographic locality calculation unit 24 performs morphological analysis on the description information, the location information, and the posted document, and generates a character string that can specify the location such as the facility name, the place name, and the address from the description information, the location information, and the posted document. Extract (step S22). Then, the geographic locality calculation unit 24 refers to the place information table and acquires position information associated with the place specified by the extracted character string (step S23). Further, the geographic locality calculation unit 24 uses the geotag as position information.

Then, the geographic locality calculation unit 24 uses the N pieces of position information x _i acquired in step S23 to calculate the reciprocal of the dispersion of the position information as shown in the equation (1), and the calculation result Is a score S1 (step S24). Then, the geographical locality calculation unit 24 outputs the calculated score S1 to the determination unit 27 together with the user ID of the target user. As described above, the geographic locality calculation process ends, and the process proceeds to step S03.

  Subsequently, the user type determination device 20 performs a formality calculation process (step S03). In the officialness calculation process in step S03, as shown in FIG. 11, first, the officialness calculation unit 25 first determines the number of followers, the number of followers, description information, and the profile associated with the user ID of the target user. The image is acquired from the SNS information DB 23 (step S31). At this time, the officialness calculation unit 25, for example, among the SNS information associated with the user ID of the target user, from the SNS information having the latest posting date and time, the number of followers, the number of followers, description information, and the profile Get images and

  Then, the officialness calculation unit 25 sets the acquired follow number as the follow number C_follow and the acquired follower number as the follower number C_follower. Then, the formula calculation unit 25 calculates the ratio Rf of the follower number C_follower to the follow number C_follower as shown in the equation (2) (step S32). And the officialness calculation part 25 calculates the logo degree Pl which shows the possibility that a profile image is a logo from a profile image (step S33). Further, the officialness calculation unit 25 calculates the facility level Ps based on the business hours, the telephone number, and the presence / absence of the mail address from the description information (step S34).

  Then, the officialness calculation unit 25 calculates the weighted sum of each value as shown in Expression (3) using the calculated follow number C_follow, follower number C_follower, ratio Rf, logo degree Pl, and facility degree Ps. Then, the calculation result is set as a score So (step S35). Then, the officialness calculation unit 25 outputs the calculated score So to the determination unit 27 together with the user ID of the target user. As described above, the formality calculation process ends, and the process proceeds to step S04.

  Subsequently, the user type determination device 20 performs a sociality calculation process (step S04). In the sociality calculation process of step S04, as shown in FIG. 12, first, the sociality calculation unit 26 acquires the posted document and the total number of posts associated with the user ID of the target user from the SNS information DB 23. (Step S41). And sociality calculation part 26 sets the total number of contributions matched with the newest contribution date and time among the acquired total number of contributions as total number of contributions Ca.

  The sociality calculation unit 26 analyzes the acquired posted document, counts the number of posted documents including the at sign “@”, and sets the counted number as the conversation number Cm (step S42). Then, as shown in the equation (4), the sociality calculation unit 26 calculates the ratio of the conversation number Cm of the target user to the total posting number Ca, and sets the calculation result as the relative conversation number Cr (step S43). .

  Then, the sociality calculation unit 26 calculates the weighted sum of each value using the calculated number of conversations Cm and the number of relative conversations Cr as shown in the equation (5), and sets the calculation result as the score Ss. (Step S44). Then, the sociality calculation unit 26 outputs the calculated score Ss to the determination unit 27 together with the user ID of the target user. As described above, the sociality calculation process ends, and the process proceeds to step S05.

  Subsequently, the user type determination device 20 performs a user determination process (step S05). In the user determination process in step S05, as illustrated in FIG. 13, first, the determination unit 27 determines whether or not the score S1 is equal to or greater than a threshold value Sth1 (step S51). When it is determined in step S51 that the score S1 is equal to or greater than the threshold value Sth1 (step S51; Yes), the determination unit 27 determines whether the score So is equal to or greater than the threshold value Sth2 (step S52).

  When it is determined in step S52 that the score So is equal to or greater than the threshold value Sth2 (step S52; Yes), the determination unit 27 determines whether the score Ss is equal to or less than the threshold value Sth3 (step S53). When it determines with score Ss being below threshold value Sth3 in step S53 (step S53; Yes), the determination part 27 determines with a target user being a regional user (facility official user) (step S54).

  On the other hand, if it is determined in step S51 that the score S1 is less than the threshold value Sth1 (step S51; No), if it is determined in step S52 that the score So is less than the threshold value Sth2 (step S52; No), and step When it is determined in S53 that the score Ss is larger than the threshold value Sth3 (step S53; No), the determination unit 27 determines that the target user is a user other than the regional user (step S55). As described above, the user determination process ends, and the process proceeds to step S06.

  Subsequently, the user type determination device 20 performs a local user information storage process (step S06). In step S06, when it is determined in step S05 that the target user is a regional user, the determination unit 27 acquires a user name and a screen name associated with the user ID of the target user from the SNS information DB 23. . Then, the determination unit 27 stores the acquired user name and screen name in the regional user information DB 28 as regional user information in association with the user ID. In step S22, when a character string that can specify the location can be extracted from the location information, the determination unit 27 further includes the location information indicating the location specified by the character string in the region user information. Stored in the user information DB 28. Then, a series of processes of the user type determination method performed by the user type determination device 20 ends.

  In the user type determination device 20 described above, based on the SNS information including the target user's profile information and document information, the score Sl regarding the target user's geographical locality, the score So regarding the target user's officialness, and the target user's A score Ss regarding sociality is calculated. Then, based on the score S1, the score So, and the score Ss, it is determined whether or not the target user is a regional user.

  Local users often post documents that describe information within a specific region. For this reason, regional users are likely to have a higher geographical locality than other users. Therefore, when the score S1 is large, it can be determined that the target user is a regional user, and the regional user can be specified. In addition, the reliability of information included in documents posted by official users is high. For this reason, it is possible to specify a regional user who transmits highly reliable information regarding a specific region by determining whether or not the target user is a regional user in consideration of officialness. Furthermore, the official user has a high possibility of not having a conversation with other users individually, and is likely to have low sociality. For this reason, it is possible to improve the accuracy of determining whether or not the user is an official user by considering sociality.

  In the above embodiment, when the score S1 is equal to or greater than the threshold Sth1, the score So is equal to or greater than the threshold Sth2, and the score Ss is equal to or less than the threshold Sth3, it is determined that the target user is a regional user. In this case, a user with high geographical locality and officialness and low sociality is determined as a regional user. Since the information transmitted by such a regional user is reliable information regarding a specific region, it is possible to identify a regional user who transmits highly reliable information regarding a specific region.

  When it is determined that the target user is a regional user, the regional user information is added to the regional user information DB 28 based on the profile information of the target user. By storing regional user information, it is possible to easily search for regional users who transmit information regarding various regions. In other words, since it is highly possible that the local user is a facility official user, it is possible to easily search for posts from the facility official user. As a result, the user of the portable terminal can easily browse the latest information (tweets) around the facility and the facility.

  As mentioned above, although embodiment of this invention was described, this invention is not limited to the said embodiment. For example, the geographical locality calculation unit 24 is not limited to the distribution of the position information acquired from the SNS information, and may calculate the score Sl using another index that can determine the geographical locality.

  Further, the officialness calculation unit 25 may calculate the score So by using at least one of the follow number C_follower, the follower number C_follower, the ratio Rf, the logo degree Pl, and the facility degree Ps, and determines the officialness of the user. The score So may be calculated by further using another obtained index. For example, since it is considered that reliable information is transmitted to a corporate user, a value indicating whether or not the user is a corporate user may be used as an index of formality. Moreover, since the authenticated user has a high possibility of transmitting reliable information, the authentication information of the SNS information may be used as an officialness index. In this case, for example, “1” may be assigned if the authentication information is “True”, and “0” may be assigned if it is “False”.

  Further, the sociality calculation unit 26 may calculate the score Ss using at least one of the conversation number Cm and the relative conversation number Cr, and further uses the other index that can determine the sociality of the user to score Ss. May be calculated.

  Moreover, the determination part 27 is good also as determining whether an object user is a facility official user among local users based on the score S1 and the score So. Furthermore, the determination unit 27 may determine whether or not the target user is a facility official user among regional users based on the score S1, the score So, and the score Ss. As this determination method, the same method as the determination method of the regional user can be used.

  The SNS user includes a bot user, which is an automatic posting system using a machine. When the bot user distributes information on a specific area, the above-described determination method may not be able to distinguish between the facility official user and the bot user. Therefore, the officialness calculation unit 25 acquires at least one of the user name and the screen name associated with the user ID of the target user from the SNS information DB 23. The formality calculation unit 25 determines whether or not the character string “bot” is included in at least one of the user name and the screen name using the regular expression, and when the character string “bot” is included. The value of the score So may be decreased (for example, zero). Thereby, it can be avoided that the bot user is determined as the facility official user.

  When the facility name can be extracted from the user name of the target user determined to be the regional user, the determination unit 27 further includes the facility name in the regional user information associated with the user ID of the target user. You may store in local user information DB28.

  DESCRIPTION OF SYMBOLS 20 ... User type determination apparatus, 22 ... Reception part (acquisition means), 24 ... Geographic locality calculation part (geographic locality calculation means), 25 ... Formality calculation part (formality calculation means), 26 ... Sociality Calculation unit (sociality calculation unit), 27... Determination unit (determination unit), 28... Regional user information DB (regional user information storage unit).

Claims (8)

A user type determination device that determines whether or not a user who uses a communication service that posts a document and views a posted document is a regional user who transmits information about a specific region,
Obtaining means for obtaining service information including profile information relating to the user's profile and document information relating to a document posted by the user;
Based on the service information, a geographical locality calculating means for calculating a first feature quantity related to the geographical locality indicating a degree of limitation of a geographical range including a place included in the information transmitted by the user;
Determination means for determining whether the user is the regional user based on the first feature amount;
A user type determination device comprising:   The geographical locality calculating unit acquires one or more pieces of position information based on the service information, and calculates the first feature amount based on a distribution of the acquired one or more pieces of position information. The user type determination device described in 1. Based on the service information, further comprising a formality calculating means for calculating a second feature amount related to the formality indicating the reliability of the information transmitted by the user;
The user type determination device according to claim 1, wherein the determination unit determines whether the user is the regional user based on the first feature value and the second feature value.   The officialness calculation means is based on the second feature based on the number of links indicating the number of other users linked to the user and the number of linked users indicating the number of other users linked to the user. The user type determination device according to claim 3, wherein the amount is calculated. Based on the service information, further comprising a sociality calculation means for calculating a third feature value related to sociality indicating the degree to which the user communicates with other users,
The said determination means determines whether the said user is the said local user based on the said 1st feature-value, the said 2nd feature-value, and the said 3rd feature-value. User type determination device.   The user according to claim 5, wherein the sociality calculating unit calculates the third feature amount based on a number of documents posted by the user toward another user among documents posted by the user. Type determination device.   When the first feature amount is greater than or equal to a first threshold, the second feature amount is greater than or equal to a second threshold, and the third feature amount is less than or equal to a third threshold, The user type determination device according to claim 5 or 6, wherein the user type determination device determines that the user is a regional user. A local user information storage means for storing local user information related to the local user;
The determination unit adds the regional user information to the regional user information storage unit based on the profile information of the user when the user determines that the user is the regional user. The user type determination device according to any one of claims 7 to 9.

Images