Message-ID: <1308522201801-3610056.post@n4.nabble.com>
Date: 2011-06-19T22:23:21Z
From: tonyp
Subject: Rolling Correlation Matrixes
In-Reply-To: <1308512396424-3609808.post@n4.nabble.com>
Hi Achim,
Thanks for your prompt reply. Sounds R is a toy for you to play around with.
mycor <- function(x) {
rval <- cor(x)
rval[lower.tri(rval)]
}
I can see from the first code that you are suggesting a function for
calculating the lower triangle of the correlation matrix and the second
seems to apply a columnwise average somehow. I just don't see how it
calculates the empirical percentile to each entry of the last(pw) matrix.
maybe I misunderstand the function?
apply(pw, 2, function(x) mean(x <= tail(x, 1)))
Can you suggest me, if there is similar function as for pairs() that can be
customized. Once I calculate the last rolling correlation matrix and the
current percentiles I want to put in the lower half what your "mycor"
calculates and in the upper half the matched percentiles for the current
rolling cor matrix in a graphical way?
If anybody else can help me with it, I would totally appreciate it. :)
Thanks again. Seems that I have to learn this tool long way.
Best wishes,
T
--
View this message in context: http://r.789695.n4.nabble.com/Rolling-Correlation-Matrixes-tp3609808p3610056.html
Sent from the Rmetrics mailing list archive at Nabble.com.