Re: [R] Detecting Growth Trends
Date: Tue, 31 Aug 2010 14:37:16 -0700 (PDT) From: rtist patwarner2...@yahoo.com To: r-help@r-project.org Subject: Re: [R] Detecting Growth Trends Message-ID: 1283290636694-2402347.p...@n4.nabble.com Content-Type: text/plain; charset=us-ascii If the test rejects the null, then it has determined that the new set of incoming data is no longer purely oscillatory in the mean reverting sense (it is now unit root and exhibits growth). Unless I misinterpreted, the OP wants to find a statistical method to determine such behavior beyond purely eyeballing the data stream. Since I assume the OP is also referring to a dynamic data stream, one could simply run the test with a sliding window and use the p result to determine if the regime is changing from mean reverting to growth. I'm not sure how an alternate form of regression fitting says anything about the question as I interpreted it (although, hopefully the poster can respond). -- Hi, No intention of starting a flamewar and thanks for the many suggestions. Indeed in my case I am not looking for a fit, but, as you nicely put it, for something a bit more on the quantitative side than eyeballing. Many thanks Lorenzo __ R-help@r-project.org 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.
[R] Detecting Growth Trends
Dear All, I am given some noisy data which (by naked eye) appears to be oscillating first but finally growing. Is there any statistical set (I mean something different from e.g. a linear fit, which would not be convincing at all in my case) to detect growth (possibly without relying on any data fitting)? Many thanks Lorenzo __ R-help@r-project.org 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] Detecting Growth Trends
You can try a unit root test which test for stationarity in a series. -- View this message in context: http://r.789695.n4.nabble.com/Detecting-Growth-Trends-tp2402080p2402132.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org 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] Detecting Growth Trends
Error of the 3rd kind, (right answer to wrong question), I think. So what if the test rejects --- then what? I think the poster is looking for some kind of smoother. ?loess, ?smooth.spline and about 400 others may be useful. -- Bert Gunter Genentech Nonclinical Statistics On Tue, Aug 31, 2010 at 11:52 AM, rtist patwarner2...@yahoo.com wrote: You can try a unit root test which test for stationarity in a series. -- View this message in context: http://r.789695.n4.nabble.com/Detecting-Growth-Trends-tp2402080p2402132.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.htmlhttp://www.r-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. [[alternative HTML version deleted]] __ R-help@r-project.org 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] Detecting Growth Trends
If the test rejects the null, then it has determined that the new set of incoming data is no longer purely oscillatory in the mean reverting sense (it is now unit root and exhibits growth). Unless I misinterpreted, the OP wants to find a statistical method to determine such behavior beyond purely eyeballing the data stream. Since I assume the OP is also referring to a dynamic data stream, one could simply run the test with a sliding window and use the p result to determine if the regime is changing from mean reverting to growth. I'm not sure how an alternate form of regression fitting says anything about the question as I interpreted it (although, hopefully the poster can respond). -- View this message in context: http://r.789695.n4.nabble.com/Detecting-Growth-Trends-tp2402080p2402347.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org 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.