A Method of Social Network Node Preference Evaluation Based on the Topology Potential

Abstract

This paper reports a hypergraph model for online social networks with an emphasis on the node preference. Some improvements of the model are made in the present study. First, the inherent nodes properties and their links are utilized in the proposed evaluation model. Second, the proposed model contains a topology potential value of node, which is based on cognitive data field in physics. In the calculation of node quality entropy - weight method are used. In way, human interference factors can be obtained for estimating node quality. The calculation of shortest path is based on the Dijkstra hypergraph. Third, a replacing algorithm is employed to account for node preference by modifying deleting algorithm. Then, based on the node preference and the feature that nodes are attracted each other in data field to form a community, we propose a hyper-graph model for the function of social networks community detection. The model is experimented to prove the validity and usability of evaluation results.