Detecting Communities and Surveying the Most Influence ofOnline Users

Social network is a virtual environment that provides services forconnecting users with the same interests, points of view, gender,space and time. Beside connection, information exchange,communication, entertainment and so on. Social network is alsoan environment for users who work in online business, advertisement or politics, criminal investigation. How to knowwhat users discuss topics via exchanged contents and communitieswhich users join in? In this paper, we propose a model by usingtopic model combined with K-means to detect communities ofonline users. Each user in social network is represented by avector in which the components are the distribution probabilitiesof interested topics of that user. Based on the components of thisvector, we discover the interested topics of online users to detectcommunities and survey users who are the most influence incommunities to recommend for spreading information on socialnetwork.

Detecting Communities and Surveying the Most Influence ofOnline Users Keywords:

LDA , ART , K-Means , online community , topic model , influence