Inferring Topic-Level Influence from Network Data

Abstract

Existing influence analysis research has largely focused on studying the maximization of influence spread in the whole network, or inferring the “hidden” network from a list of observations. There is little work on topic-level specific influence analysis. Although some works try to address this problem, their methods depend on known social network structure, and do not consider temporal factor which plays an important role in determining the degree of influence. In this paper, we take into account the temporal factor to infer the influential strength between users at topic-level. Our approach does not require the underlying network structure to be known. We propose a guided hierarchical LDA approach to automatically identify topics without using any structural information. We then construct the topic-level influence network incorporating the temporal factor to infer the influential strength among the users for each topic. Experimental results on two real world datasets demonstrate the effectiveness of our method. Further, we show that the proposed topic-level influence network can improve the precision of user behavior prediction and is useful for influence maximization.