Topic based object Influence analysis in Social networks

Abstract

In this article, we focus on researching models in social networks analysis and use them to develop the system which enables users to exactly find topics as well as necessary research of authors. In this research, we study Author – Conference – Topic (ACT) model to find out the topic relevant to users’ requirements. With each of these topics, the system will create a graph of Topical Affinity Propagation (TAP) in purpose of analyzing the topic based influence of authors. Experimental results on the corpus extracted from Microsoft Academic Research include 34,330 authors, 19,921 articles, 1,335 conference  and 424,501 co-author relations.