ESTIMASI REGRESI WAVELET THRESHOLDING DENGAN METODE BOOTSTRAP

Abstract

Wavelet is a function that has the certainly characteristic for example, it oscillate about zero point ascillating, localized in the time and frequency domain and construct the orthogonal bases in L2(R) space. On of the wavelet application is to estimate non parametric regression function. There are two kinds of wavelet estimator, i.e., linear and non linear wavelet estimator. The non linear wavelet estimator is called a thresholding wavelet rstimator. The application of the bootstrap methode in the thresholding wavelet function estimation is resample the wavelet coefficient of residual. The best of the thresholding wavelet estimator with bootstrap method has minimal of mean square error (MSE). The minimal MSE depend from the number of replication.