基于小波分解—支持向量机的短时交通量预测

基于交通流预测问题与函数估计和逼近问题是等价的的思想,提出一种基于小波分解-支持向量回归的短时交通量预测方法。首先对交通量数据进行小波分解,然后分别对基本信号和不同分辨率的干扰信号建立支持向量机模型,最后对多个预测结果进行合成,从而得到交通量的预测结果,并利用实例计算显示模型具有较低的误差,证明了该方法具有很好的可靠性。

Short-term Traffic Forecast Based on WD and SVM

According to the equivalence of traffic forecasting and function estimating and approaching,the essay puts forward a method of short-term traffic forecasting which is based upon wave decompounding and support vector regress.The method first analyses the wave decompound traffic data,sets up support vector machine model separately on basic signals and disturbing signals of different resolving rate,then compounds several forecasting results,finally the traffic forecasting result is obtained.With examples,the result shows that the model has relatively lower error rate and quite reliable.