An improved method based on the recursive shingling algorithm

Identifying communities in large-scale social networks is not only a complex social computing problem but also a research hotspot. Furthermore, in real social networks, people often keen to find contact-intensive, interactive frequently dense communities. In this view, the paper analyzes a traditional method of community detection, the recursive shingling algorithm based on similarity measures. And on this basis, we presented an improved method, and identify that the improvement method is better than the recursive shingling algorithm through the community density, the strength of community and the proportion of important nodes experiments.