A Grid-Based Approach for Similarity Mining of Massive Geospatial Trajectories

With the proliferation of mobile devices accompanied by the advancement in location detection technologies, such as GPS, GSM network logs, call description records (CDR), etc., a large amount of spatio-temporal data is being generated. A lot of research is directed towards understanding and discovering cumulative movement and behaviour patterns of users from this digital footprint. An interesting problem in this category is to be able to identify similarity between mobile users' trajectories. In this paper, a generic grid-based approach for user similarity mining from such location logs is presented. The technique is applied on real geo-location data to derive trajectory similarity patterns.