Open AccessPENERAPAN METODE WAVELET THRESHOLDING UNTUK MENGAPROKSIMASI FUNGSI NONLINIER
Author(s) -
Muhammad Luthfie Janariah,
Syamsul Bahri,
Nurul Fitriyani
Publication year - 2021
Publication title -
indonesian physical review
Language(s) - English
Resource type - Journals
ISSN - 2614-7904
DOI - 10.29303/ipr.v4i3.98
Subject(s) - thresholding , wavelet , haar wavelet , mathematics , pattern recognition (psychology) , artificial intelligence , function (biology) , wavelet transform , discrete wavelet transform , computer science , algorithm , image (mathematics) , evolutionary biology , biology
The wavelet thresholding method is an approximation method by reducing noise, which is known as the denoising process. This denoising process will remove noise while closed the important information in the data. In this research, the wavelet thresholding method is used to approximate the nonlinear function. The data used for the simulation is a representation of several functions that represent several events that often occur in the real world, which consists of the types of functions Blocks, Bumps, Doppler, and HeaviSine. Based on simulation results based on the indicator value of the Cross-Validation (CV), the best approximation of the nonlinear function using the wavelet thresholding method for the four simulation cases are: (i) the Blocks function is given by Haar wavelet with a soft of thresholding function and the 10-th resolution level ; (ii) the Doppler function is given on the 2-nd order of Symlets wavelet with a soft of thresholding function and the 10-th resolution level; (iii) the Bumps function is given on the 6-th order of Daubechies wavelet with a soft of thresholding function and the 10-th resolution level; and (iv) the HeaviSine function is given by the 3-rd order of Coiflet wavelet with a soft of thresholding function and the 7-th resolution level.