A new combinatorial algorithm based on the characteristics of wavelet transform, particle swarm optimization algorithm and BP neural network for short-term traffic flow forecasting was presented in this paper. Firstly, using wavelet transform, we do multi-resolution decomposition of traffic flow data and single branch reconstruction on its original scale, then apply the particle swarm optimized BP neural network to forecast on each reconstruction sequence, at last, add all the forecasting results to the final short-term traffic flow forecasting results. The experimental results show that, comparing to the traditional particle swarm optimized BP neural network and traditional BP neural network, the new algorithm significantly improves the forecasting accuracy of short-term traffic flow, has a broad application prospect and is more suitable for forecasting short-time traffic flow data which contains much more noise or has a serrated signal curve.
Short-term traffic flow forecasting based on wavelet transform and neural network
2017-10-01
220535 byte
Conference paper
Electronic Resource
English
The Forecasting Approach for Short-term Traffic Flow Based on Wavelet Analysis and Neural Network
British Library Online Contents | 2007
|Wavelet-Chaotic Integration-Based Forecasting for Short-Term Traffic Flow
British Library Conference Proceedings | 2007
|A Wavelet Network Model for Short-Term Traffic Volume Forecasting
Online Contents | 2006
|