| Re: canonical correlations [message #51300] |
Wed, 15 November 2006 05:27 |
dvila
Messages: 13
Registered: April 2006
|
Junior Member |
|
|
Jared,
Canonical correlation is a general case the product moment correlation.
In canonical correlation analysis you measures the relationship
between the observed values of two *set* of variables (instead of two
variable with linear corralation). I was wondering if someone knows an
IDL code to perform this issue.
Thanks again!
Daniel
kuyper@wizard.net wrote:
> Jared Espley wrote:
>> Hi Daniel,
>>
>> I'm not sure what exactly you mean by "canonical" but the 'correlate'
>
> http://en.wikipedia.org/wiki/Canonical_correlation
|
|
|
|
|
|
| Re: canonical correlations [message #51308 is a reply to message #51307] |
Tue, 14 November 2006 12:47  |
Jared Espley
Messages: 17
Registered: February 2006
|
Junior Member |
|
|
Hi Daniel,
I'm not sure what exactly you mean by "canonical" but the 'correlate'
function gives the Pearson correlation and 'r_correlate' does rank
correlations such as Spearman's and Kendall's.
See the IDL help file and http://www.nrbook.com/a/ for more info on
these topics.
Good luck,
Jared
dvila wrote:
> Hi all,
>
> Do you know if are there any available IDL code to perform canonical
> correlations/analysis? Thank you very much!
>
> Daniel Vila
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
