Wavelet significance testing with respect to GWN background: Monte Carlo simulation usage

Abstract

The paper deals with significance testing of wavelet coefficients. We investigate the test of wavelet power spectrum with respect to the Gaussian white noise background spectrum from two perspectives of calculating significance level: with the use of χ2 distribution and with the use of Monte Carlo simulation. Our experiment is performed on a special kind of synthetic signal in which the frequency component is changing during time. We investigate the level of variance σ2 of the signal and the significance risk α. We describe the advantages and disadvantages of both approaches and formulate recommendations for using time-frequency testing.