EBM: an entropy-based model to infer social strength from spatiotemporal data

摘要: The ubiquity of mobile devices and the popularity location-based-services have generated, for first time, rich datasets people's location information at a very high fidelity. These can be used to study behavior - example, social studies shown that people, who are seen together frequently same place most probably socially related. In this paper, we interested in inferring these connections by analyzing information, which is useful variety application domains from sales marketing intelligence analysis. particular, propose an entropy-based model (EBM) not only infers but also estimates strength co-occurrences space time. We examine two independent ways: diversity weighted frequency, through contribute strength. addition, take characteristics each into consideration order compensate cases where limited available. conducted extensive sets experiments with real-world including both data their connections, latter as ground-truth verify results applying our approach former. show outperforms competitors.