|
From: | Muhali |
Subject: | Re: Confidence-intervall for correlation? |
Date: | Wed, 19 Oct 2011 09:03:10 -0700 (PDT) |
I don't think it is a vague request. Correlation coefficient estimates have a well known approximative distribution via the Fisher transform, see here e.g. A Matlab implementation is here.
The confidence interval is a very important diagnostic that should be part of octave (easy to implement).
M.
[Prev in Thread] | Current Thread | [Next in Thread] |