Exploring the Impact of False Location Identification on the Inference of Social Ties in Location-Based Social Networks

Global Positioning System (GPS) and Wifi facilitate the users to add extra location information to their social networks, which forms Location-Based Social Networks (LBSNs). Analysis of social networks formed over LBSNs are helpful in understanding human behavior and is utilized in e-business recommender systems. Co-occurrences between two sets of users based on various locations give us a clue about the social connection between the two users. But sometimes, locations like Bus Stop and public areas give us the wrong indication of friendship. In our research, we treated such a set of locations as False Locations. For improved social ties inference, we propose a framework, that takes the LBSNs as an input and generates two sets of results. The first result is a set of inferred friendship relations that is based on mutual friend and co-occurrence count. The second result is a set of friendship ties based on mutual friend and co-occurrence count and False Location. We tested the performance of the proposed framework using Gowalla and CDR datasets. Results identify those public places have a negative impact on social ties inferring.