Social Context-Based Non-overlapping Communities' Detection Model in Social Networks

In recent years, several models, approaches, and algorithms for analyzing and extracting knowledge from social networks (SN) have been proposed. One of the most desired knowledge in this context is finding a grouping of subscribers in clusters; we talk then about the social concept of "communities." The detection of communities has become an important task in understanding how the structure of SN changes over time. It is also an essential step in SN analysis. However, finding effectively non-overlapping communities in real data sets remains a challenge and an area of topical research that attracts many researchers. In this paper, we propose a new approach to cover the best partition of non-overlapping communities according to the number of nodes in common between each pair of them. Our experiments on a real SN show that the approach proposed can precisely define non-overlapping communities, on the one hand, and can significantly improve the quality of community detection and obtain accurate community structure, on the other hand.