MOOD: Moving objects outlier detection

Abstract

This paper describes and demonstrates MOOD, a system for detecting outliers from moving objects data. In particular, we demonstrate a continuous distance-based outlier detection approach for moving objects' data streams. We assume that the moving objects are uncertain, as the state of a moving object can not be known precisely, and this uncertainty is given by the Gaussian distribution. The MOOD system provides an interface which takes moving objects' states streams and some parameters as input and continuously produces the distance-based outliers along with some graphs comparing the efficiency and accuracy of the underlying algorithms. © 2014 Springer International Publishing Switzerland.