Human Contact Prediction Using Contact Graph Inference

Predicting human mobility is considered as a challenging problem. In this paper, we formulate the problem of human contact prediction as a graph inference problem. We show the importance of using offline social information for predicting people's contacts motivated by homophily theory. We also prove that by using the small-world network properties of the contact graphs, we can reconstruct the missing part of a contact graph where only part of the graph is known. Our results are promising because they allow researchers to reconstruct the missing parts in experimentally measured human mobility traces when only partial traces are obtainable.