COMPUTER SOFTWARE AND COMPUTING EQUIPMENT

Research of various models of wavelet filters at the analysis of time series

N. V. Demich, O. V. Demich

Astrakhan State Technical University

Abstract: Aspects of the application of wavelet models and scaling filters for the analysis of time series are considered in the paper. The analysis of properties of Haar filters and Daubechies filters is resulted. The choice of an optimum set of filters is proved at the application of wavelet-methodology for the most accurate approximation of the model of the financial market to the reality and forecasting of its statuses.

Keywords: discrete wavelet transform, discrete Fourier transform, Haar wavelet filter, Daubechies wavelet filter.

UDC: 681.3

Received: 18.11.2009