Real-time Monitoring of Mobile User using Trajectory Data Mining

This paper examines the mobility patterns of a mobile User through trajectory data mining. The spatio-temporal trajectory of a mobile User is captured with the aid of GPS devices. An efficient data dependent approach is presented to conveniently monitor in real-time, the positions of a mobile User in a specific location. Considering a spatial set of positions, P with a specified weight, w and a rectangle with specified-size, R
_i,j
. A trajectory monitoring problem is introduced, that is to identify a centroid of P to represent the centre of R
_i,j
. Such that there exists a maximised weighted-sum of P in R
_i,j
. As a result, an efficient algorithm which is termed as Trajectory Maximised Weighted-Sum (TMaxWS) has been proposed to identify any evolving P, and to monitor the trajectory of P in order to locate the centroid. The propose approach improves on the limitations of an existing approach including G2 algorithm and Best Region Search that are based on Maximised Range-Sum (MaxRS) without localisation and recognition of a mobile User's position and location data. In this paper, the propose approach makes use of a grid-based index structure, data dependent multi-object classes and branch and bound formalism. The experimental evaluation with Geolife as real-life massive dataset has proven that TMaxWS can adequately localise and recognise the positions and location data of a mobile User. It performs relatively faster on update of valid and efficient results.