混沌时间序列的局域线性回归预测方法
A Local Regression Forecasting Method of Chaotic Time Series
Abstract
混沌时间序列预测是80年代末发展起来的一种非线性预测新方法.它已在天气预报、经济预测、电力负荷预测、股市预测等方面得到成功应用.混沌运动是确定系统具有内在随机性的一种运动,它的行为极其敏感地依赖于初始条件.混沌系统从两个极其邻近的初始点出发的两条轨道... In order to improve the effectiveness of chaotic time series forecasting, a local regression forecasting method based on chaos theory is proposed. The effectiveness of the proposed method is demonstrated by Lorenz chaotic time series and an time series of annual average sunspot numbers.