Skip to content

autocorrelation test

2 messages · Poizot Emmanuel, Spencer Graves

Original
Poizot Emmanuel
# 
Hi all,

I would like to test the relevance of a vector field (i.e. if the 
vectors are organized or not).
To do so, I would like to use an autocorrelation test, so that I have 
two questions:
- is the Watson test applicable to that perpuse ?
- is the kuiper test applicable to that purpuse ?

Regards

------------------------------------------------
Emmanuel Poizot
Cnam/Intechmer
B.P. 324
50103 Cherbourg Cedex

Phone (Direct) : (00 33)(0)233887342
Fax : (00 33)(0)233887339
------------------------------------------------
3 days later
Spencer Graves
# 
I'm not certain I understand your question.  However, I did the 
following as I thought it might produce something that you could use:

	  > RSiteSearch("vector field correlation")

	  This produced 25 hits, several of which might interest you.  In 
particular, I wonder of you could use the function "loglik.GRF in the 
geoR package?  This computes "Log-Likelihood for a Gaussian Random 
Field" 
(http://finzi.psych.upenn.edu/R/library/geoR/html/loglik.GRF.html).  If 
you fit two nested models and compute this for both, then you can 
compute from these numbers 2*log(likelihood ratio).  Standard asymptotic 
theory says that in many situations, this number will be be distributed 
approximately chi-square with degrees of freedom equal to the difference 
in the number of parameters estimated in the two models.

	  If you'd like further help from this list, please submit another 
question.  Before you do, however, I suggest you read the posting guide! 
"www.R-project.org/posting-guide.html".  Anecdotal evidence suggests 
that posts more consistent with this guide tend to receive quicker, more 
useful replies.

	  hope this helps.
	  spencer graves
Poizot Emmanuel wrote: