基于在线LS-SVM算法的变参数混沌时间序列预测

研究利用最小二乘支持向量机(LS-SVM)预测变参数混沌时间序列。变参数混沌系统适合于描述现实中的复杂混沌现象,但由于参数的慢变导致系统动力学特性不断发生变化,基于Tan-kens嵌入定理的建模预测方法难以适用,其时间序列预测可以看作是小样本学习问题。最小二乘支持向量机是在二次损失函数下采用等式约束求解问题的一种支持向量机,保留支持向量机优点同时计算量大大减少。提出用一种具有遗忘机制的最小二乘支持向量机在线递推算法,并引入历史数据的高次项预测变参数混沌时间序列。对典型变参数混沌时间序列的预测结果表明,该方法具有较高预测精度,能快速跟踪预测变参数混沌时间序列。

Predict the Time Series of the Parameter-varying Chaotic System Based on Recursive Lease Square Support Vector Machine(RLS-SVM)

The forecast on the time series of the parameter-varying chaotic system was researched using LS-SVM in this paper. The parameter-varying chaotic system is fit for illustrating the practically complicated chaotic phenomena,and some of these systems′ dynamical character always changes with their parameter′s slow shifting. So the forecast models based on the Takens embedding theory can not be applied on such system,especially the prediction on their time series can be regard as the classical learning problem o...