高嵌入维混沌负荷序列预测方法研究

现有的采用欧氏距离确定相空间最邻近点的混沌预测方法对高维混沌时间序列预测的效果不太理想,因而提出以关联度代替欧氏距离来确定相空间最邻近点的思想,同时发展了一种改善高嵌入维重构空间全局Lyapunov指数谱性状的方法.通过对短期电力负荷序列的预测,验证了当嵌入维数逐渐增大时,所提方法比现有的方法在预测精度方面有明显的提高.

RESEARCH OF FORECASTING METHOD ON CHAOTIC LOAD SERIES WITH HIGH EMBEDDED DIMENSION

The forecasted results of high dimension chaotic time series by existing chaotic forecasting method, in which the nearest points in phase space are determined by Euclid distance, are not perfect, therefore, a thought is put forward in which to determine the nearest points in phase space the Euclid distance is replaced by correlation degree, at the same time a method is developed by which the character of full Lyapunov exponential spectrum in reconstruction space with high dimensions is improved. The forecasting results of short term power load series by the presented method show that when the embedded dimension increased gradually, the result of load series forecasting by the presented method is far more accurate than that by the existing method.