混沌时间序列局域预测方法

在深入研究混沌时间序列局域预测方法的基础上,提出了一种加权局域基函数预测方法。该方法综合考虑了广义自由度和邻近点权重,提出了加权动态确定最邻近点数的判定条件,并利用基函数拟合确定出的最邻近点进行预测。算例分析表明,加权局域基函数法具有较高的预测精度,是比较理想的用于混沌时间序列的预测方法。

Local prediction research of chaotic time series

The prediction method of weight local basis function is presented based on the deep research on local prediction for chaotic time series. Based on the theory of phase space reconstruction, with considering the generalized degrees of freedom and the neighbors' weight an improved criterion for selection of the optimal neighborhood is proposed. The prediction is realized by approximating neighbor points of the neighborhood with fitting of basis function. It means that the current point's motion is replaced by the neighbors' ones. The calculation results indicate the proposed method has superior predictive precision, and it is a much ideal local prediction method for chaotic time series.