A Numerical Investigation of Various Forms of Wavelet in Financial Time Series Analysis
Kittikorn Sriwichai, Panu Sam-ang, Sayan Kaennakham · 2021 International Conference on Electrical, Computer and Energy Technologies (ICECET) · 2021
In this work, seven forms of discrete wavelet transformation under the mother wavelet type are numerically studied for analyzing the financial time series data. They are Haar, Daubechies, Discrete FIR approximation of Meyer wavelet, Symlets, Coiflets, Biorthogonal, and Reverse biorthogonal. The data used consist of the Dow Jones Index (DJIA 30) from 17 July 2000 until 16 July 2020. Our numerical investigation shows that Haar performs best in solving the autocorrelation problem and denoising data.