Title: Community detection on social network with complex attributes
Authors: Wanjie Wang - National University of Singapore (Singapore) [presenting]
Abstract: Social network analysis gains more and more interest in current studies, together with the case that nodes also have attributes. There are many studies on how to combine the network data and attribute data to achieve better community detection results, yet most of them are for low-dimensional attributes. In the statistician citation network data, we take the paper abstracts as the attribute data, which is high-dimensional and sparse. We proposed a new community detection method for such data with complex attributes, based on the SCORE method. We apply the method to the network data and get interesting results. We also prove corresponding theoretical results on the algorithm.