Hi,
I would like to fit a linear model where some but not all explanators are chosen
stepwise - ie I definitely want to include some terms, but others only if they
are deemed significant (by AIC or whatever other approach is available). For
example if I wanted to definitely include x1 and x2, but only include z1 and z2
if they are significant, something like this:
df <- data.frame(y=c(4,2,6,7,3,9,5,7,6,2), x1=c(2,3,4,0,5,8,8,1,1,2),
x2=c(0,0,0,0,1,1,0,0,0,1), z1=c(0,1,0,0,0,1,1,0,1,1), z2=c(1,1,1,0,0,1,1,1,1,0))
model <- lm(y ~ x1 + x2 + stepwise(z1 + z2))
Any help would be appreciated.
Cheers,
Andre
**********************************************************************
This email and any attachments are confidential, protect...{{dropped:22}}
Resist the temptation. Stepwise analysis without shrinkage will ruin model inferences without helping with predictive accuracy. Frank AndreE wrote:> > Hi, > > I would like to fit a linear model where some but not all explanators are > chosen stepwise - ie I definitely want to include some terms, but others > only if they are deemed significant (by AIC or whatever other approach is > available). For example if I wanted to definitely include x1 and x2, but > only include z1 and z2 if they are significant, something like this: > > df <- data.frame(y=c(4,2,6,7,3,9,5,7,6,2), x1=c(2,3,4,0,5,8,8,1,1,2), > x2=c(0,0,0,0,1,1,0,0,0,1), z1=c(0,1,0,0,0,1,1,0,1,1), > z2=c(1,1,1,0,0,1,1,1,1,0)) > model <- lm(y ~ x1 + x2 + stepwise(z1 + z2)) > > Any help would be appreciated. > > Cheers, > Andre > ********************************************************************** > This email and any attachments are confidential, protect...{{dropped:22}} > > ______________________________________________ > R-help@ mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide > http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. >----- Frank Harrell Department of Biostatistics, Vanderbilt University -- View this message in context: http://r.789695.n4.nabble.com/Select-some-but-not-all-variables-stepwise-tp3990002p3990026.html Sent from the R help mailing list archive at Nabble.com.
Francisco Mora Ardila
2011-Nov-04 13:20 UTC
[R] Select some, but not all, variables stepwise
Hi Andre I don?t know if it will work, but I?ve tried the MuMIn package, were you can evaluate all possible models (usin for example AIC) at one time. Maybe you can focus on comparing those models which retain the explanators you want. Best wishes Francisco On Fri, 4 Nov 2011 13:06:09 +0000, Andre Easom wrote> Hi, > > I would like to fit a linear model where some but not all explanators are > chosen stepwise - ie I definitely want to include some terms, but others only > if they are deemed significant (by AIC or whatever other approach is available) > . For example if I wanted to definitely include x1 and x2, but only include > z1 and z2 if they are significant, something like this: > > df <- data.frame(y=c(4,2,6,7,3,9,5,7,6,2), x1=c(2,3,4,0,5,8,8,1,1,2), x2=c(0,0, > 0,0,1,1,0,0,0,1), z1=c(0,1,0,0,0,1,1,0,1,1), z2=c(1,1,1,0,0,1,1,1,1,0)) model > <- lm(y ~ x1 + x2 + stepwise(z1 + z2)) > > Any help would be appreciated. > > Cheers, > Andre > ********************************************************************** > This email and any attachments are confidential, protect...{{dropped:22}} > > ______________________________________________ > R-help at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.Francisco Mora Ardila Laboratorio de Biodiversidad y Funcionamiento del Ecosistema Centro de Investigaciones en Ecosistemas UNAM-Campus Morelia Tel 3222777 ext. 42621 Morelia , MIchoac?n, M?xico. -- Open WebMail Project (http://openwebmail.org)