Community detection in location-based social networks: An entropy-based approach

Community detection is a fundamental problem in social network analysis. There are a large number of community detection algorithms in the literature. However, many previous algorithms only focus on the structure of the network, and ignore the attributes of the users (e.g., the location attributes of the users). In this paper, we study the community detection problem in the context of location-based social networks. We propose a new location-aware community detection algorithm, called LBLPA, based on a new entropy-based similarity and a label propagation procedure. Specifically, in our algorithm, we first make use of the frequency of users' POI (point of interest) to represent the location features, and then calculate the similarity between users by a novel entropy-based method. Subsequently, we propose a label propagation algorithm (LBLPA) to detect the community based on such a newly proposed entropy-based similarity measure. We also propose a location-aware metric, called Structural Entropy, to evaluate the quality of the detected communities. Finally, we conduct extensive experiments on three real-world datasets, and the results demonstrate the effectiveness and efficiency of our algorithms.