You can use predict() by specifying a complete data.frame() for prediction to
the argument newdata=. Look:
da <- expand.grid(x1=LETTERS[1:4], x2=1:9)
da$y <- rnorm(da$x1)
da$y[sample(length(da$y), 5)] <- NA
m0 <- lm(y~x1+x2, data=da)
predict(m0) # NA not predicted
predict(m0, newdata=da) # NA pr
see ?na.exclude
you can set na.action='na.exclude' when fit the model.
On 15 April 2010 09:06, Martin Batholdy wrote:
> Hi,
>
> I wanted to use the predict.lm() function to compare the empirical data with
> the predicted values.
> The problem is that I have NAs in my data.
>
> I wanted to cbin
Hi,
I wanted to use the predict.lm() function to compare the empirical data with
the predicted values.
The problem is that I have NAs in my data.
I wanted to cbind my data.frame with the empirical values with the vector I get
from predict.lm.
But they don't have the same length because predict.
3 matches
Mail list logo
