Mobility Intention-Based Relationship Inference From Spatiotemporal Data

Inferring social relationship based on co-occurrence has become a focal point in the last decade. Many studies indicate that more frequently two users co-occur at non-public locations, the higher probability they are acquaintances. We find that in some spatiotemporal datasets collected by Internet of Things (IOT) devices in public locations, it's hard to distinguish co-occurrences between acquaintances and strangers. In this paper, we propose a mobility intention-based relationship inference model (MIRI) to address above challenge. We utilize mobility intention to characterize co-occurrences and propose a classification model for social relationship inference. The experimental results on real-world dataset demonstrate not only the superiority of our model, but also improve the effectiveness.