Spatio-Temporal Data Mining for Location-Based Services

Location–Based Services (LBS) are continuously gaining popularity. Innovative LBSes integrate knowledge about the users into the service. Such knowledge can be derived by analyzing the location data of users. Such data contain two unique dimensions, space and time, which need to be analyzed. The objectives of the thesis are three–fold. First, to extend popular data mining methods to the spatio–temporal domain. Second, to demonstrate the usefulness of the extended methods and the derived knowledge in promising LBS examples. Finally, to eliminate privacy concerns in connection with spatio–temporal data mining by devising systems for privacy–preserving location data collection/mining. To this extent, first, a general methodology, pivoting, is described to extend a popular data mining method, namely rule mining, to the spatio–temporal domain. By considering the characteristics of a number of real–world data sources, a taxonomy of spatio–temporal data is derived, and the usefulness of the rules that the extended spatio–temporal rule mining method can discover is demonstrated. The proposed spatio–temporal extension is applied to find long, sharable patterns in trajectories of moving objects. Empirical evaluations show that the extended method and its variants, using high–level SQL implementations, are effective tools for analyzing trajectories of moving objects. Since real–world trajectory data about a large population of objects moving is difficult to obtain, to aid the development in spatio–temporal data management and data mining, next the development of a Spatio–Temporal ACTivity Simulator (ST–ACTS) is described. ST–ACTS uses a number of real– world geo–statistical data sources and intuitive principles to effectively generate realistic spatio–temporal activities of mobile users. Premium SMS Cab-Sharing Service Geocoding Service