基于小波分析的月径流ARIMA预测方法

Wavelet Analysis-based ARIMA Method for Monthly Runoff Forecast

将一种基于小波分析的自回归滑动平均求和(ARIMA)模型用于月径流的预测。首先利用小波变换良好的局部化特性,将月径流序列分解成不同时间尺度上的子序列;然后对各个子序列利用ARIMA模型进行预测。将采用基于小波分析的ARIMA模型的预测结果与直接使用ARIMA模型的预测结果进行比较,结果表明引入小波变换提高了月径流预报精度。

A wavelet analysis-based ARIMA model is proposed to forecast monthly runoff. With the help of the localization characteristic of wavelet transform, the monthly runoff series is first decomposed to sub-series on different time scales, and each sub-series is modeled and forecasted by the ARIMA model. Then, the forecast result from the wavelet analysis-based ARIMA model is compared with the result from the single ARIMA model. It is shown that the introduction of wavelet transform helps improve the precision of monthly runoff forecast.