Travel time prediction on urban networks based on combining rough set with support vector machine

Urban travel time prediction is one of interesting topics in current transportation research and practice. In this paper, a new prediction model is proposed which combines rough set with support vector machine. Rough set is used to pre-process the traffic data that is noisy, missing, and inconsistent then deduce some rules for framing support vector machine (SVM) model. When comparing the committee model to the single SVM predictions utilizing real traffic data collected in Chengdu, it is concluded the new approach indeed leads to improved travel time predicting accuracy and velocity.