Data mining for predicting traffic congestion and its application to Spanish data

Abstract

The purpose of this paper is the development and application of patterns and behavioral models of time series data collected by sensors belonging to the Spanish Directorate General for Traffic. The extraction of these patterns will be used to predict the behavior and effects on the system as accurately as possible to facilitate early notifications of traffic congestions, therefore minimizing the response time and providing alternatives to the circulation of vehicles. Decision trees, artificial neural networks and nearest neighbors algorithms have been successfully applied to a particular location in Sevilla, Spain.