论文部分内容阅读
为获得更精确的径流-水位预报结果,利用Dmey小波变换将水位时间序列分解为高频信号和低频信号,再使用均生函数-最优子集回归对其进行预测,最后利用Dmey小波逆变进行重构,以此建立水位预测模型。通过对柳江历年水位进行实例分析,并与均生函数-最优子集回归模型、逐步回归模型对比。研究结果表明,该模型能充分反映水位时间序列趋势,预报稳定性好,预报准确率高,为径流-水位时间序列预测提供一个有效建模方法。
In order to obtain more accurate results of runoff-water level prediction, Dmey wavelet transform is used to decompose the water level time series into high-frequency and low-frequency signals, and then predicted by means of the mean-squared-optimal subset regression. Finally, Dmey wavelet transform Reconstruction, in order to establish water level prediction model. Through the case analysis of Liujiang River water level over the years, and with the mean of life function - the optimal subset regression model, stepwise regression model comparison. The results show that the model can fully reflect the time series trend of water level, the forecasting stability is good, and the prediction accuracy rate is high, providing an effective modeling method for runoff - water level time series forecasting.