Label propagation method based on constraint about triangles for community detection in complex networks

Label propagation is a strategy for optimizing some objects to detect community structure. In this paper, we introduce a new object about triangles to guide the label update. To avoid the trend that merges too many nodes into a large community, we add one constrain on the objective. Through the experiments and comparison, we select the suitable strength for the constraint. Last, we merge the objective and the existing one linearly into a hybrid objective and use the hybrid objective to guide the label update in our proposed label propagation algorithm. We perform amounts of experiments on artificial and real-world networks. Experimental results demonstrate the superiority of our algorithm in both modularity and speed.