Partitioning signed bipartite graphs for classification of individuals and organizations

In this paper, we use signed bipartite graphs to model opinions expressed by one type of entities (e.g., individuals, organizations) about another (e.g., political issues, religious beliefs), and based on the strength of that opinion, partition both types of entities into two clusters. The clustering is done in such a way that support for the second type of entity by the first within a cluster is high and across the cluster is low. We develop an automated partitioning tool that can be used to classify individuals and/or organizations into two disjoint groups based on their beliefs, practices and expressed opinions.