Denoising Malaysian time series data: A comparison using discrete and stationary wavelet transforms

Wavelets are designed to comprise certain properties that would make them a useful mathematical tool for signal processing. One application of discrete wavelet transform (DWT) is in analyzing financial time series data. The purpose of this paper is to apply DWT and stationary (discrete) wavelet transform (SWT), namely Haar, Daubechies, Symmlet and Coiflet in denoising a financial time series data from Kuala Lumpur Stock Exchange (KLSE) and compare the results amongst the four wavelets. The data consists of 4056 daily data of closing index starting from December 3, 1993 until May 7, 2010. The results show that Daubechies wavelet produced a better approximation of the data compared to the Haar, Symmlet and Coiflet wavelets.