[R] predict newdata question

[Bill.Venables at csiro.au]

You ask:

# How can I use predict here, 'newdata' crashes
predict(m1,newdata=wolf$predicted);wolf # it doesn't work

To use predict() you need to give a fitted model object (here m1) and a *data frame* to specify the values of the predictors for which you want predictions.  Here wolf$predicted is not a data frame, it is a vector.

What I think you want is

pv <- predict(m1, newdata = wolf)

That will get you linear predictors.  To get probabilities you need to say so as

probs <- predict(m1, newdata = wolf, type = "response")

You can put these back into the data frame if you wish, e.g.

wolf <- within(wold, {
	lpreds <- predict(m1, wolf)
	probs <- predict(m1, wolf, type = "response")
})

Now if you look at 

head(wolf)

you will see two extra columns.
 

-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf Of Felipe Carrillo
Sent: Saturday, 26 June 2010 10:35 AM
To: r-help at stat.math.ethz.ch
Subject: [R] predict newdata question

Hi:
I am using a subset of the below dataset to predict PRED_SUIT for
the whole dataset but I am having trouble with 'newdata'. The model
was created with 153 records and want to predict for 208 records. 

[lots of stuff omitted]

wolf$prob99<-(exp(wolf$predicted))/(1+exp(wolf$predicted))
head(wolf);dim(wolf)
   # How can I use predict here, 'newdata' crashes
 predict(m1,newdata=wolf$predicted);wolf  # it doesn't work

Thanks for any hints


Felipe D. Carrillo
Supervisory Fishery Biologist
Department of the Interior
US Fish & Wildlife Service
California, USA




______________________________________________
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.