[R] partial autocorrelation plots ACF type=p

Tonker claire_brittain at hotmail.com
Wed Oct 8 15:25:29 CEST 2008


Dear users,

I have two continuous variables which are two different measures taken each
year from 1975 to 2005. I want to see if the two variables are correlated
but need to take into account the fact that they are a time series. I have
been following an example from 'The R Book' where you plot the ACF: 
par(mfrow=c(1,1)
acf(cbind(x,y))

and this appeared to work fine, producing four graphs for the
autocorrelation of x,x y,y x,y and y,x.

Then the next step in the example was to plot the partial autocorrelation:
acf(cbind(x,y),type="p")

When i do this however three of my graphs x,x y,y and y,x contain no data,
the x,y graph contains data but has no lines of significance and all four of
the graphs are on vertical scales of -2000 to 6000.

I am new to time series analysis and from what i have read the partial
autocorrelation plots should come out on a vertical scale of -1 to 1. I dont
know why my partial autocorrelation plots are not working when the
autocorrelation plots seemed to be no problem. If anyone has any suggestions
I would be grateful.

Thank you in advance.
Claire
-- 
View this message in context: http://www.nabble.com/partial-autocorrelation-plots-ACF-type%3Dp-tp19878665p19878665.html
Sent from the R help mailing list archive at Nabble.com.



More information about the R-help mailing list