中国物理B ›› 2009, Vol. 18 ›› Issue (8): 3287-3294.doi: 10.1088/1674-1056/18/8/032
闫华, 魏平, 肖先赐
Yan Hua(闫华), Wei Ping(魏平), and Xiao Xian-Ci(肖先赐)
摘要: Due to the error in the measured value of the initial state and the sensitive dependence on initial conditions of chaotic dynamical systems, the error of chaotic time series prediction increases with the prediction step. This paper provides a method to improve the prediction precision by adjusting the predicted value in the course of iteration according to the evolution information of small intervals on the left and right sides of the predicted value. The adjusted predicted result is a non-trajectory which can provide a better prediction performance than the usual result based on the trajectory. Numerical simulations of two typical chaotic maps demonstrate its effectiveness. When the prediction step gets relatively larger, the effect is more pronounced.
中图分类号: (Time series analysis)