A System for Traffic Events Detection Using Fuzzy C-Means

Abstract

Systems for traffic events administration are important tools in the prediction of disasters and management of that of the movement flow in diverse contexts. These systems are generally developed on non-fuzzy grouping algorithms and ontologies. However, the results of the implementation do not always give high precision scores due to different factors such as data heterogeneity, the high number of components used in their architecture and to the mixture of highly specialized and diverse domain ontologies. These factors do not ease the implementation of the systems able to predict with higher reliability traffic events. In this work, we design a system for traffic events detection that implements a new ontology called trafficstore and leverages the fuzzy c-means algorithm. The indexes evaluated on the fuzzy c-means algorithm demonstrates that the implemented system improves its efficiency in the grouping of traffic events.