Stable Label Propagation Algorithm Based on Edge Weights for Community Detection

Lei ZHANG,Feng QIAN,Shu ZHAO,Jie CHEN,Yan-ping ZHANG,Feng LIU

DOI: https://doi.org/10.3969/j.issn.1000-1220.2019.02.014

2019-01-01

Abstract:In order to avoid the instability of the output result and the defect of the giant community formation in the label propagation algorithm (LPA), a Stable Label Propagation Algorithm based on Edge Weights (SLPA_EW) for community detection is proposed.First, we measure the different statuses of node neighbors based on the triangle structure and uses the metrics as the edge weights.Next, the edge weights of the same label in the neighbors are used to guide the updating process of the node labels and the random label selection is changed to definitized selection to ensure the stability of the output of the algorithm. In order to avoid the label oscillation problem in the process of label initialization, take the strategy of forming node pairs and giving the same label to these nodes. In the process of label propagation, adding label weight to the label updating strategy limits the scale of community formation and avoids the problem of giant community. The experimental results on real network datasets showthat the proposed algorithm can guarantee the stability and quality of the output and have lowrunning times and iteration times.