Research on Short-term Traffic Flow Prediction Method Based on CS Optimized RVM
The core content of the intelligent transportation system is short-term traffic flow prediction,so the improvement of traffic flow prediction accuracy and the reduction of prediction time are the key issues of current research.Aimed at this prob-lem,a short-term traffic flow prediction model based on Cuckoo search(CS)algorithm optimized relevance vector machine(RVM)regression is proposed.The traffic flow data are denoised and normalized based on the PeMS.The road occupancy ratio and average speed are used as the input and the traffic flow is used as the output to be the training set in RVM prediction model.It finds the optimal value of the kernel width parameter σ in the RVM by the CS algorithm to improve the performance of the algorithm and obtain the best prediction model.Compared with the RVM and support vector regression(SVR)predic-tion models,the proposed CS-RVM traffic flow prediction model has higher prediction accuracy and shorter time.
万方数据
维普
