Prediction of the chaotic time series using multivariate local polynomial regression

Abstract

摘要 根据Takens定理，把混沌时间序列构造为一组序列对，然后用多元局部多项式方法来预测其序列.这种核估计方法可以结合局域法与全局法的优点，使得预测的精度更高.仿真结果表明，该方法非常有效. 关键词: 混沌时间序列 / 多元局部多项式方法 / 核估计 Abstract According to the Takens theorem, we reconstruct the chaotic time series to multivariate data and use the multivariate local polynomial estimator to predict the chaotic time series. This kernel estimator has the advantages of both the local model and the global model. The simulation shows that the estimator can predict the chaotic time series accurately. Keywords: chaotic time series / multivariate local polynomial regression / kernel estimator 作者及机构信息 周永道, 马 洪, 吕王勇, 王会琦 基金项目: 国家自然科学基金(批准号:10571127)和高等学校博士学科点专项科研基金(批准号：20040610004)资助的课题. Authors and contacts 参考文献 施引文献