基于小波变换的交通流预测及模型优化

Wavelet Transform Based Traffic Flow Predicting and Model Optimization

将小波变换的理论应用于短时交通流量的预测。通过小波分解与重构获取交通流量数据中的低频近似部分和高频随机部分,然后在分析各种模型的优、劣的基础上,选取较有效的模型或模型结合方式,建立了交通流量预测模型。最后,利用实测交通流量数据对模型仿真,结果表明该模型可以有效地提高短时交通流量预测的精度。同时,深入研究了小波基、分解层数选取的依据,通过实验检验表明效果良好。

The theory is applied of wavelet transform into short-time traffic flow predicting. Initially. The traffic flow data via wavelet decomposition and reconstruction respectively can be obtain approximate and detail. Next, we can compare different kinds of models to get effective forecasting model then establish the model of traffic flow predicting. Finally, a simulation with online measured original data indicates a broad prospect on improving the precision of short-time traffic flow predicting with this model. Simultaneously, an intensive study is made of on the Selection criterion of wavelet base and the number of decomposing layers, which was justified by experiment.