Increasing the accuracy of identifying overlapping communities using weighted edges

In recent years, several algorithms have been proposed for detection in complex networks. It should be noted that wuth regarding the features of these communities, one of the existing methods for identifying communities, is providing an algorithms for weighting the edges of the network as the weight of the edges of the communities are incresed and at the same time the weight of the edges between communities are reduced. Therefore the distinction between communities could be identified simply. In the proposed method with useing the process of weighting the edges, we distinguish between the nodes that are more similar to each other and the nodes that have slight similarity. i.e. by assigning the weight with using the proposed criteria in some algorithms, the edges with more weight will have a greater role in determining the population.According that there is a positive correlation between similarity measures and community structures, the results of the tests shows that using local similarity measures as the weight of edges for some algorithms, causes an increase in the accuracy of communities recognition. These algorithms use the degree of the nodes as one of the network characteristics for computing the cores absorbing ability for communities formation. For example, in the case of Real Networks, running WHD-EM algorithm on the High school network, discovers the communities with the NMI=0.6652 and the purity criteria equal to 0.9845. Also it should be noted that some algorithms such as CPM, OSLOM and LINK in terms of NMI criteria, are better.