When is the periodogram is consistent with white noise?
A programme I wrote in R could be relevant. The reference is I. Lobato and C. Velasco, Econometric Theory, Vol.20, 2004, "A Simple and General Test for White Noise". The acf function and the spec.pgram function are used to produce a transformed von Mises statistic which is approx. N(0,4). Tests are standard and finite sample values are easily generated. The programme seems to work OK in experiments using random numbers and AR1 series. If there is any interest I can supply a script file.
Andre Bastos wrote:
Hello everyone, This is my first time posting to the list, thanks in advance. I am calculating the smoothed periodogram for the residuals of an AR model that I fit to EEG data. The autocorrelation plot of the residuals shows the series is now approximately white (i.e. ACF = 1 at lag 0, and close to 0 for all other lags). I would like to show that the spectrum of the series is also approximately white. When I calculate the periodogram using spec.pgram, what I get indeed looks white, but I would like to add horizontal threshold lines between which one can be 95% confident the spectrum is white. Thanks, -Andre [[alternative HTML version deleted]]
______________________________________________ R-help at stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
View this message in context: http://www.nabble.com/When-is-the-periodogram-is-consistent-with-white-noise--tf4050952.html#a12930061 Sent from the R help mailing list archive at Nabble.com.