Modeling POI Transition Network of Human Mobility

摘要

Point-of-interest (POI) denotes the category of a location people visit, reflecting the daily life and personal interests. Through a massive dataset of check-in records, this paper discloses that the POI visiting pattern of human spatial mobility falling into both Zipf's law and Heaps' law. It is found that people always have three potential choices when selecting a particular POI as their next travel destination, namely, exploring a new POI, returning to a visited one, or staying at the current one. We calculate the probabilities of these scenarios and propose a statistical model based on random walk to represent such phenomena. Simulation results show that our model can well reproduce the heterogeneous Zipf's law and Heaps' law of human mobility.