[R] Partial whitening of time series?
I have a time series with a one year lag, ar=0.5. The series has some interesting events that disappear when the series is whitened (i.e., fitting an AR process and looking at the residuals). I'd like to remove the autocorrelation in stages to see the effect on the time series. Is there a way to specify the autocorrelation term while fitting an AR process? For instance, given the following: x - arima.sim(model = list(order = c(1,0,0), ar = 0.5), n = 500, sd=0.25) Can I filter x in a way that the autocorrelation at lag one is 0.4, then 0.3, 0.2, 0.1, until I get to a clean series equivalent to: y - arima(x, order = c(1,0,0))$resid Thanks in advance, Andy __ R-help@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.
Re: [R] Partial whitening of time series?
andy, if your model is Xt = 0.5 * Xt-1 + e, then it should have Xt = 0.1 * Xt-1 + 0.4 * Xt-1 + e (Xt - 0.1*Xt-1) = 0.4 * Xt-1 + e so what you need to do is to substract part of lag from your series. it is just my $0.02. On 2/26/07, Andy Bunn [EMAIL PROTECTED] wrote: I have a time series with a one year lag, ar=0.5. The series has some interesting events that disappear when the series is whitened (i.e., fitting an AR process and looking at the residuals). I'd like to remove the autocorrelation in stages to see the effect on the time series. Is there a way to specify the autocorrelation term while fitting an AR process? For instance, given the following: x - arima.sim(model = list(order = c(1,0,0), ar = 0.5), n = 500, sd=0.25) Can I filter x in a way that the autocorrelation at lag one is 0.4, then 0.3, 0.2, 0.1, until I get to a clean series equivalent to: y - arima(x, order = c(1,0,0))$resid Thanks in advance, Andy __ R-help@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. -- WenSui Liu A lousy statistician who happens to know a little programming (http://spaces.msn.com/statcompute/blog) __ R-help@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.
Re: [R] Partial whitening of time series?
Thanks, I wasn't thinking real clearly when I pressed 'send'. All figured out now. -A -Original Message- From: Wensui Liu [mailto:[EMAIL PROTECTED] Sent: Monday, February 26, 2007 10:15 AM To: Andy Bunn Cc: r-help@stat.math.ethz.ch Subject: Re: [R] Partial whitening of time series? andy, if your model is Xt = 0.5 * Xt-1 + e, then it should have Xt = 0.1 * Xt-1 + 0.4 * Xt-1 + e (Xt - 0.1*Xt-1) = 0.4 * Xt-1 + e so what you need to do is to substract part of lag from your series. it is just my $0.02. On 2/26/07, Andy Bunn [EMAIL PROTECTED] wrote: I have a time series with a one year lag, ar=0.5. The series has some interesting events that disappear when the series is whitened (i.e., fitting an AR process and looking at the residuals). I'd like to remove the autocorrelation in stages to see the effect on the time series. Is there a way to specify the autocorrelation term while fitting an AR process? For instance, given the following: x - arima.sim(model = list(order = c(1,0,0), ar = 0.5), n = 500, sd=0.25) Can I filter x in a way that the autocorrelation at lag one is 0.4, then 0.3, 0.2, 0.1, until I get to a clean series equivalent to: y - arima(x, order = c(1,0,0))$resid Thanks in advance, Andy __ R-help@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. -- WenSui Liu A lousy statistician who happens to know a little programming (http://spaces.msn.com/statcompute/blog) __ R-help@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.