Community discovery is a crucial task in social network analysis, especially in describing the evolution of social networks. Although some works have focused on finding the dynamic community, there are still some open problems need to be conquered, such as analyzing the dynamic and weighted community. In this paper, we propose a framework for analyzing weighted communities and their evolutions via clustering correlated weight vectors to enhance existing community detection algorithms. The International trade network is used to verify our framework. Experiments show that the framework discovers and captures the evolving behaviors with temporal elements and weight values.
Research and Practice of Personalized Learning Recommendation Based on Dynamic Community Discovery
