论文部分内容阅读
针对径流量时间序列非线性、非平稳性的特点,利用经验模态分解法将其分解为多个不同频率下的时间序列组合揭示演变规律,采用双向差分法对分解后的各序列逐个反导其微分方程,并结合自忆性原理构建预测模型,由叠加分解后的各序列预测值获得径流量的预测值。实例结果表明,该方法充分挖掘了数据自身信息,拟合效果较好,预测年径流量准确,具有推广应用价值。
According to the characteristics of non-linear and non-stationary runoff time series, the empirical mode decomposition method was used to decompose it into several time-series combinations under different frequencies to reveal the evolution law. Two-way difference method was used to decompose the decomposed sequences one by one The differential equation and the self-remembering principle are used to construct the predictive model, and the prediction values of the runoff are obtained from the predicted values of each sequence after the superposition. The results of the example show that this method fully exploits the information of the data, and has a good fitting effect and accurate prediction of annual runoff, which is of great value in popularization and application.