Research And Applicationon Of Temporal Network Node Similarity

With the development and progress of Internet technology,as a Cross-discipline with great application value,network science has gradually become the hotspot of domestic and foreign scholars.With the research of network science getting deeper,the demand for its applications has gradually become diversified.Temporal network research has also become an important research direction in complex network analysis.Because temporal network research can effectively retain the temporal information contained in the network,it is essential to discover the hidden information in the complex structure of the network.As an important measure in complex network research,node similarity is the basis of node relation research.At present,its research is mainly confined to the case of static network,difficult to meet the application requirements.Changes in real network structure makes the node similarity is also constantly changing.The similarity of nodes in static networks is actually difficult to distinguish the similarity between nodes.As the hotspot of social network research the research of community detection also faces the same problem in recent years.And the local community discovery algorithm with temporal information will become an inevitable trend.This paper starts from the research background of temporal network similarity,introduced the research status of related fields.Elaboratedthe definition and characteristics of temporal network,several representative temporal network models,the static network node similarity measure and the classical community discovery algorithm.Based on this,the temporal network model is established,the similarity measure of temporal network and temporal local modularity is defined.A local community discovery algorithm for temporal networks is designed.In the experiment stage,a series of contrastive experiments are designed using temporal network data set.By comparing the performance in link prediction,the superiority of similarity measure of temporal network nodes is verified.The stability of the algorithm is proved by choosing different initial nodes in global community detection.The experiment results show that the node similarity measure based on temporal network is more reasonable.The proposed community detection algorithm has important theoretical and practical value.