ABSTRACT
Testing a signal to see if it is stationary (or not) has attracted many researchers’ interests. (Dwivedi & Rao 2011) proposed a method to test second order stationarity based on discrete Fourier transform according to the fact the discrete Fourier transform is asymptotically uncorrelated at the canonical frequencies if and if only the time series is second order stationary. (Brcich & Iskander 2006) developed a statistical test for stationarity based on the ratio of the arithmetic to geometric means of spectra obtained from non-overlapping segments of the time series. (Steven 2008) using a Rao test to determine the stationarity length of a locally wide sense stationary Gaussian random process by modeling the process as a time-varying autoregressive (TVAR) model. (Basu 2009) proposed a nonparametric test for stationarity by checking that the statistics of transform coefficients over epochs of the signal do not deviate greatly from their sample mean and two test statistics based on the periodogram and multitaper estimators of the power spectral density and studied their sampling distribution under the null hypothesis.
