Investigation of the role of similarity measure and ranking algorithm in mining social networks

Social networks have attracted much attention recently. Different studies have been conducted to automatically extract social networks among various kinds of entities from the web. Social network analysis finds its application in many current business areas. In this paper we demonstrate how the choice of the similarity measure affects ranking results of entities in a social network extracted from the web. We use different similarity measures in order to build different social networks. By applying formulas described below for each of the networks we derive a new network which is different from the original one by edge weights. Subsequently, in the derived networks we rank entities again. Finally we compare the results.