Abstract: There is provided methods and systems for analyzing data from a plurality of users within a social networking platform, comprising: receiving a query for a topic associated with the social networking platform; determining a set of users having at least one social networking behavior on the social networking platform related to the topic; selecting, for each user from the set of users, a pre-defined number of posts and associating each of the pre-defined number of posts with the respective user; segmenting the selected posts for each user to determine a likelihood of each of the selected posts among the set of users; and, clustering the selected posts for each user to define a plurality of clusters and determining a mapping from each user to at least one of the plurality of clusters, each cluster comprising representative topics indicating frequently used topics within the cluster for the pre-defined number of posts between the set of users.
Type:
Grant
Filed:
November 4, 2014
Date of Patent:
June 14, 2016
Assignee:
SYSOMOS L.P.
Inventors:
Brian Jia-Lee Keng, Edward Dong-Jin Kim