Trajectories are spatiotemporal data generated by moving objects containing the spatial position of object at various time intervals. GPS devices record this information and it is possible to construct trajectory of moving objects for analysis. Outlier analysis of trajectories is done to identify abnormal activities like intrusion detection, fraud detection, fault detection and rate event detection. In this paper, Trajectory Outlier Detection algorithm using Boundary (TODB) is proposed using a boundary construction algorithm and a binary classifier. In TODB, Convex Hull algorithm is used to construct the boundary and ray casting algorithm is used to build the binary classifier. TODB is tested for its accuracy using real world data sets. Experimental results on real world data sets demonstrate that TODB correctly classify normal and outlier trajectories.
cited By 0
B. A. Sabarish, Karthi, R., and Kumar, T. G., “Spatial Outlier Detection Algorithm for Trajectory Data”, International Journal of Pure and Applied Mathematics, vol. 118, pp. 325-330, 2018.