Traffic Volume Prediction Algorithm Based on Traffic Flow Sequence Partition and a Neural Network

Abstract

It is an important premise to identify and predict traffic flow instantly and accurately in ITS. Therefore, it is of great significance to realize the control and induction of ITS. Aiming at the intersection short-term traffic volume forecasting problem, we proposed the combined model prediction algorithm based on the analysis of traffic flow sequence partition and neural network model. This algorithm divides the traffic volume into different patterns along the time and volume dimension by clustering analysis, and then describes and predicts traffic flow value according to different patterns. The experiment results on real data set demonstrate that our algorithm based on the combination model is more accurate.