Traffic Status Prediction and Analysis Based on Mining Frequent Subgraph Patterns

Abstract

With the development of the city, the traffic congestion and traffic accidents on the urban road increase frequently. Using traffic modeling and analysis to improve the traffic conditions become more important. Now, using the traffic flow model to study the traffic problems has made many achievements. However, traffic flow model cannot be a good choice for describing the relations of the traffic element at a specific moment, but these relations are indeed significant for forecasting traffic status from that moment on. In this paper, a graph model for the static traffic was studied, and then analyzed the feature of a graph substructure for traffic congestion at one moment. We propose an effective frequent subgraph mining algorithm to find the frequent substructure that represent traffic congestion status in a graph. Our mining algorithm can enhance the efficiency of finding the congestion subgraph. Analyzing the proportion of the congestion subgraph in a graph for traffic to forecast the traffic status at that moment later, thus to find ways to improve traffic conditions.