Time related association rules mining for traffic prediction based on genetic network programming combined with estimation of distribution algorithms

In this paper, a method of time-related class association rule mining is proposed based on Genetic Network Programming (GNP) combined with Estimation of Distribution Algorithms (EDAs). There are two important points in this paper: The first important point is to combine GNP with Estimation of Distribution Algorithms which are a novel evolution strategy. The second important point is that three kinds of probability models have been put forward for generating new individuals. The aim of this paper is to extract more interesting association rules and to improve the traffic prediction accuracy by combining Genetic Network Programming with Estimation of Distribution Algorithms. We applied the proposed data mining algorithm to traffic systems in order to predict the traffic volume in future. The simulation results show that our proposed method is effective compared with the conventional method based on GNP.