Short Term Diverging Traffic Flow Prediction based on ImF-VOMM Method

Abstract

Traffic congestion brings many problems in human daily life, and further leading to environment pollution and economic losses. The more precise prediction of traffic status is, the better we can settle the problems. The prediction of traffic flow is more valuable than traffic speed because of its characteristic of directivity. In this paper, in order to further improve predicting accuracy, we utilize the conception of association rule to discover regions that have strong correlation to neighboring regions. Then adding the neighboring regions information into the original regions diverging flow prediction. On the basis of Variable-order Markov Model (VOMM), integrating the correlation information as the impact factor, we finally process a novel predicting method IF-VOMM method. Experimental results show that our method could improve the predicting accuracy and have more stability.