Entropy-based Social Network Link Partition Algorithm

摘要

It is an important core issue in social networks to divide community or group. And, the network node is the mainstream of community division algorithm as the processing object to divide the network. This paper introduces the entropy theory into social networks partition on the basic of studying the concept of entropy and social networks partition algorithm. And we proposed an Entropy-based Link Partition algorithm (ELP algorithm), which is the social network links as the processing object. Also, the similarity between two objects is properly defined and improved, which thus is more close to the real situation of the social network. Experimentation on two real-world networks, and we obtained results of community division and compared with other community partition algorithms to verify the effectiveness of the proposed algorithm. The ELP algorithm has a higher accuracy, and communities are more realistic than that generated by either of the Link Clustering algorithm (LC) or the classical Clique Percolation Method (CMP).