[R] bootstrap validation of LR error message

Marc Schwartz marc_schwartz at comcast.net
Fri Jan 16 21:37:19 CET 2009


on 01/16/2009 02:19 PM A Van Dyke wrote:
> when i try to validate my logistic regression model:
> 
> fit<-glm(y~x,binomial,data=dataname,x=TRUE,y=TRUE)
> validate(fit,method="boot",B=150,...)
> 
> i get the following error message:
> 
> Error in UseMethod("validate") : no applicable method for "validate"
> 
> any insight would be appreciated.  many thanks!


You appear to be trying to use the validate() function from Frank
Harrell's Design package.

However, you are attempting to use it with a glm() created from R's
stats package. The two don't mix.

If you want to use Frank's validate() function, you will need to create
your logistic regression model using Frank's lrm() function.

See ?lrm and ?validate.lrm for more information.

HTH,

Marc Schwartz




More information about the R-help mailing list