[19282] in s-news-athena
RE: [S] Studentized residuals from nls
daemon@ATHENA.MIT.EDU (Wouters, Luc - 2744 [JanBe])
Wed Sep 22 04:32:11 1999
Message-Id: <27CD61B9B111D211B0DC0008C72435B7FF1073@janbebeex1.janbe.jnj.com>
From: "Wouters, Luc - 2744 [JanBe]" <lwouters@janbe.jnj.com>
To: "Laffont, Jean-Louis" <Laffont@phibred.com>,
"'s-news@wubios.wustl.edu'" <s-news@wubios.wustl.edu>
Date: Wed, 22 Sep 1999 10:17:06 +0200
Mime-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Dear Jean-Louis,
The following function should do the job:
"StudRes.nls"<-function(grad, nls.fit)
{
# Studentized residuals for nls
# grad: derivative of model as obtained from function deriv
# nls.fit: fit obtained from nls
v <- attributes(grad)$gradient
h <- diag(v %*% solve(crossprod(v)) %*% t(v))
s <- summary(nls.fit)$sigma
r <- nls.fit$residuals/(s * sqrt(1 - h))
r
}
_____________________________________________________________
Luc Wouters
Life Sciences Dept. 2744
Janssen Research Foundation Tel: +32-(0) 14-60.27.44
Turnhoutseweg 30 Fax: +32-(0) 14-60.57.88
B2340 Beerse, Belgium E-mail: lwouters@janbe.jnj.com
_____________________________________________________________
-----Original Message-----
From: Laffont, Jean-Louis [mailto:Laffont@phibred.com]
Sent: Tuesday, 21 September 1999 21:47
To: 's-news@wubios.wustl.edu'
Cc: Laffont, Jean-Louis
Subject: [S] Studentized residuals from nls
Dear S-Plus users,
Does anybody have written a function to compute studentized residuals from a
non linear model (using nls)?
Thanks,
Jean-Louis (laffont@phibred.com)
Pioneer Genetique
Aussonne, France
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
