Community Detection Based On Local Phase Synchronization On Complex Networks And Its Application

In pace with the rapid development of internet technology,there have been many large-scale social?economic?information and biological networks.The network topology and dynamics in networks attract the interest of scholars in various fields,thus gradually formed the complex network theories.A improved synchronization model could be used for coumunity detection via locking the local synchronization(phase locking)at a stable state.The detailed contents and main results are summarized as follow:1.Detecting community structure by local phase synchronization.With the assumption that the synchronization on complex networks could be divided into local and collective process,A improved Kuramoto model that could suppress collective synchronization and lock local synchronization is proposed.Simulations on artificial and real networks show the size of the communities is more in line with reality and nodes in the same community have similar phases and nodes' phases across different communities vary much.2.The model is extended to overlapping communities.We perform secondary division by calculating the state of nodes in the coumunity got through the first process of synchronization,thus the results are more accurate.The method has better performance than traditional methods with the measure of NMI(normalized mutual information)?Accuracy and FI-score.3.Applications of overlapping community detection.Firstly,we briefly introd-uce the applications of overlapping community detection on typical scenes such as accurate recommendation?road navigation?identifying zombie user?hot spot predict-ion.Then we empirically analysis the applications of overlapping community detect-ion algorithm based on local phase synchronization proposed in this paper on user re-commendation and stock market.