MANAGING UNCERTAINTY IN WEB-BASED SOCIAL NETWORKS

Abstract

Identifying key members from web-based social networks assists in assessing the risk of criminal network formation. To manage the uncertainty in complex web-based social networks, we first formally defined the binary relation and uncertainty of pages in web-based social networks. Secondly, we proposed an effective algorithm for Mining Key member from uncertain web-based social networks, called MiKey, by integrating uncertainty of pages into three centrality measures including degree, betweenness, and closeness. MiKey takes into a full consideration of the uncertainty in web-based social networks by computing the transition probability from one page to another. Furthermore, we briefly introduced the approach of calculating the k-order transition matrix of pages. Finally, we conducted experiments on real web data and the results show that MiKey is effective in discovering key pages from web-based social networks with less time deficiency than the centrality measures based algorithm.