Which Topic Will You Follow?

Deqing Yang, Yanghua Xiao, Bo Xu, Hanghang Tong, Wei Wang, Sheng Huang

Abstract: Who are the most appropriate candidates to receive a call-for-paper or call-for-participation? What session topics should we propose for a conference of next year? To answer these questions, we need to precisely predict research topics of authors. In this paper, we build a MLR (Multiple Logistic Regression) model to predict the topic-following behavior of an author. By empirical studies, we find that social influence and homophily are two fundamental driving forces of topic diffusion in SCN (Scientific Collaboration Network). Hence, we build the model upon the explanatory variables representing above two driving forces. Extensive experimental results show that our model can consistently achieves good predicting performance. Such results are independent of the tested topics and significantly better than that of state-of-the-art competitor.