基于小波分析的公路旅客运输量预测

Forecast of transport volume of highway passenger based on wavelet analysis

为了更加准确地预测动态变化的公路客运量,提出了基于小波分析的客运量的预测方法。针对公路客运量具有拟周期性、非平稳性和非线性的特点,首先利用小波变换对公路客运量时间序列进行小波分解,再重构低频信号提取其变化趋势,然后采用Antisymmetric(whole-point)方法对变化趋势进行延拓,对延拓后信号进行采样得到预测结果。实验表明,该方法可以有效地减小数据预测的误差,优于一般的曲线拟合法。

This paper introduces a forecast method based on wavelet analysis in order to obtain a more precise prediction of the varying traffic. The solution to the quasi-periodicity, non-stationary and non-linear owned by highway passenger transport volume involves deducting time series based on wavelet transform, rebuilding the low frequency signal, and extending the trend by using Antisymmetric （wholepoint） method to were obtain the forecasted results using the sampled signal. Experimental results show that the method can evidently decrease error of prediction and prove better adapted to the economic forecast than curve fitting.