Method, computer program and computer for detecting communities in social media
View Patent ↗The present invention provides at least a method includes: extracting a plurality of partial communities from a plurality of users, based on the relationships of companion messages; computing a first degree of similarity for showing the similarity of the companion partial communities, based on the relationship of a user belonging to one partial community with a user belonging to the other partial community, from among the plurality of communities; computing a second degree of similarity for showing the similarity of companion partial communities, based on words within the messages sent by users belonging to both partial communities and under the condition that the first similarity be higher than a predetermined first threshold value; and creating an integrated community by integrating the companion partial communities under the condition that the second similarity be higher than a predetermined second threshold value.
1. A computer-implemented method for clustering a plurality of users in social media, wherein the plurality of users each send messages, the computer-implemented method comprising the steps of:
extracting a plurality of partial communities from the plurality of users, wherein the plurality of partial communities are based on relationships of companion messages, which are connected by sending, replying, or forwarding between users in the plurality of users;
computing a first degree of similarity for showing a similarity of companion partial communities, wherein the companion partial communities are based on the relationship of users with other users, and wherein the first degree of similarity is based on a relationship of a user belonging to a first partial community with a user belonging to a second partial community, and wherein the relationship for the user belonging to the first partial community and the user belonging to the second partial community is determined based on information in the users' profiles, and wherein the first degree of similarity is modified based on location information;
computing a second degree of similarity for showing a similarity of companion partial communities, wherein the second degree of similarity is based on words in the messages sent by users belonging to the first and second partial communities and so that the first degree of similarity is higher than a predetermined first threshold value; and
creating an integrated community by integrating the companion partial communities so that the second degree of similarity is higher than a predetermined second threshold value.
2. The computer-implemented method according to claim 1 , wherein:
the messages include other messages sent by other users in response to a single message received from a single user; and
extracting the plurality of partial communities from the plurality of users based on whether the companion messages correspond to said single message from the single user and other messages received in response to said single message.
3. The computer-implemented method according to claim 1 , wherein each of the plurality of partial communities extracted in the step for extracting is a strong connected component.
4. The computer-implemented method according to claim 1 , wherein:
the social media stores user profile information; and
computing the first degree of similarity is based on the relationship between the profile information of a user belonging to the first partial community and the profile information of a user belonging to the second partial community.
5. The computer-implemented method according to claim 1 , wherein computing the second degree of similarity is based on whether a characteristic word in a message sent by a user belonging to the first partial community is similar to the characteristic word within a message sent by a user belonging to the second partial community.
6. The computer-implemented method according to claim 5 , wherein the characteristic word is extracted by creating feature vectors for the message.
7. The computer-implemented method according to claim 1 , wherein the messages are sampled under a prescribed condition from messages posted on the social media.
8. The computer-implemented method according to claim 1 , wherein the messages are clustered so that a prescribed keyword from the messages is posted on the social media in a prescribed time period.
9. The computer-implemented method according to claim 7 , wherein:
a posting computer is connected through a network to a clustering computer that clusters the plurality of users; and
the clustering computer receives messages sent from the posting computer in response to a prescribed condition request from the clustering computer.
10. The computer-implemented method according to claim 9 , further comprising storing the received messages in a memory of the clustering computer.
11. The computer-implemented method according to claim 1 , wherein the social media is a microblog.
12. The computer-implemented method according to claim 1 , further comprising outputting the integrated community by using a graphical user interface.
13. A non-transitory computer readable storage medium tangibly embodying a computer readable program code having computer readable instructions which, when implemented, cause a computer device to carry out a method for clustering a plurality of users in social media, wherein the plurality of users each send messages, the method comprising:
extracting a plurality of partial communities from the plurality of users, wherein the plurality of partial communities are based on relationships of companion messages, which are connected by sending, replying, or forwarding between users in the plurality of users;
computing a first degree of similarity for showing a similarity of companion partial communities, wherein the companion partial communities are based on the relationship of users with other users, and wherein the first degree of similarity is based on a relationship of a user belonging to a first partial community with a user belonging to a second partial community, and wherein the relationship for the user belonging to the first partial community and the user belonging to the second partial community is determined based on information in the users' profiles, and wherein the first degree of similarity is modified based on location information;
computing a second degree of similarity for showing a similarity of companion partial communities, wherein the second degree of similarity is based on words in the messages sent by users belonging to the first and second partial communities and so that the first degree of similarity is higher than a predetermined first threshold value; and
creating an integrated community by integrating the companion partial communities so that the second degree of similarity is higher than a predetermined second threshold value.
14. The computer-implemented method according to claim 1 , wherein location information includes country and administrative district.
15. The computer-implemented method according to claim 8 , wherein the messages comprise reply messages.
16. The computer-implemented method according to claim 8 , wherein the messages comprise resent messages.
17. The computer-implemented method according to claim 7 , wherein data for the messages conforming to the prescribed condition is received concurrently with information in the users' profiles.