[R] partial autocorrelation for Rt vs. Nt-1, ......., Nt-h

Prof Brian Ripley ripley at stats.ox.ac.uk
Wed Mar 3 13:12:58 CET 2004


Your subject line is `partial autocorrelation', but that is not what you 
describe in the body of the email.

Partial correlations are by definition between two variables, just like 
correlations (but they can be partial-ed on other variables).

Please read the posting guide and the references there, and try to `ask a 
good question' that unambiguously tells your readers what you want to 
know, rather than leave them guessing.

On Wed, 3 Mar 2004, Scott Newey wrote:

> Dear list, following a previous querry we are still stuck!
> 
> As pointed out by Erin Hodges the "ts" library includes a PACF function
> which reports the partial correlation of population density at time t
> against lagged population density.

Rather, it computes the `partial autocorrelation' that your subject line 
mentions.

> However, what we are trying to calculate is the partial correlation between
> rate of population change, Rt=log Nt/Nt-1, against lagged population
> densities. Which is the partial rate correlation function propsed by
> Berryman & Turchin (1999).
> 
> The list archives give show methods for estimating partial correlation for 2
> or 3 variables only. Could any list member please suggest how this could be
> done for more variables, say Nt-1 to Nt-10.
> 
> Thank you,
> Scott
> =============================================
> 
> Hello Tomas!
> 
> There are functions for pacf and plot.acf.
> 
> They are in library(ts)
> 
> Hope this helps!
> 
> Sincerely,
> Erin Hodgess
> Associate Professor
> Department of Computer and Mathematical Sciences
> University of Houston - Downtown
> mailto: hodgess at gator.uhd.edu
> 
> To: R-help at stat.math.ethz.ch
> Date: Sat, 28 Feb 2004 19:52:33 +0100
> Subject: [R] Stepwise regression and partial correlation for wildlife census
> 	time series
> 
> Dear List;
> 
> We are doing a time series analysis of wildlife census data. We use a
> stepwise regression of the annual per capita rate of increase against
> pervious years population size (log transformed) as suggested by
> Berryman & Turchin (2001, Oikos 92:265-270).
> 
> How can we obtain the partial correlation coefficients in R to make a
> plot of them against the lag as in a standard PACF?
> 
> Yours Sincerely
> 
> Tomas Willebrand
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
> 
> 

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595




More information about the R-help mailing list