An Efficient Algorithm for Detecting Communities in a Tripartite Networks

An efficient Algorithm is proposed for community detection through combining collaborative filtering and modularity maximization. We design the algorithm with divide and conquer thinking, i.e. divide a large network into smaller sub-networks and then conquer finding communities of each sub-network. In the first part of our Algorithm, we use sparse matrix technology for storage and processing to improve filtration efficiency. In the second part, a quantity filter coefficient is defined to degrade the time complexity of finding the leading eigenvector of matrix in modularity maximization method. We demonstrate the method with applications to a tripartite network and get a relatively ideal result.