Personal destination pattern analysis with applications to mobile advertising

Abstract

Many researchers expect mobile advertising to be the killer application in mobile business. In this paper, we introduce a trajectory prediction algorithm called personal destination pattern analysis (P-DPA) to analyse the different destinations in various trajectories of an individual, and to predict a trajectory or a set of destinations that could be visited by that individual. The P-DPA algorithm works on an individual level. Every destination-pattern analysis is related to the self-history and the personal profile of a targeted individual, not on what others do. In addition, we developed a prototype system called SmartShopper. SmartShopper is a personal destination-pattern-aware pervasive system for mobile advertising in (outdoor and indoor) retail environments. The predicted destinations from the P-DPA algorithm will be used by SmartShopper to generate a list of relevant advertisements adapted to the personal profile of previous destinations of a targeted individual. We tested the destination prediction accuracy of the P-DPA algorithm with a synthetic dataset of a virtual mall and a real GPS dataset.