Earticle

Most of the local community detection algorithms based on node similarity often simply count the number of common neighbors as the basis of selecting members of the community that cannot accurately measure the value of a common neighbor node in the information transmission of nodes. For this, we use the new concept of a common neighbor contribution; borrowed from the idea about the local modularity, put forward a new fast community detection algorithm. The algorithm accurately selects candidate nodes to join the community, according to the contribution of the common neighbor node, also without calculating local modularity for each common neighbor node, and greatly improved the accuracy and efficiency in merging Members. the experimental results of the computer-generated network and the real networks verified reliability and efficiency of the algorithm.