Wavelet Based Estimate for Non-Linear and Non-Stationary Auto-Regressive Model
Tôru Fujii, 亨 藤井, Takashi Yanagawa, 堯 柳川 · Kyushu University Institutional Repository (QIR) (Kyushu University) · 2003
The wavelet estimator of regression function in a non-linear auto-regressive model for non-stationary time series data is proposed. A convergence theorem of the estimators, and also related theorems, is developed. Cross-Validation criterion is proposed for the optimum selection of parameters. The criterion is justified by the $ alpha $-mixing condition. Finally, the method is applied to the electroencephalograph (EEG) data.