R help - Feb 2006 - how to extract predicted values from a quantreg fit?

Hi,

I have used package quantreg to estimate a non-linear fit to the  
lowest part of my data points. It works great, by the way.

But I'd like to extract the predicted values. The help for  
predict.qss1 indicates this:

predict.qss1(object, newdata, ...)
and states that newdata is a data frame describing the observations  
at which prediction is to be made.

I used the same technique I used to extract predicted values of GAM  
fits with mgcv package: if I fit a GAM using x as the x variable on  
which I smooth, I then create such a dataframe like this

MyX <- data.frame(x=seq(-1,60))

This works fine with GAM (mgcv) but not with quantreg:

 > y <- rnorm(500, 10, 5)
 > x <- rep(seq(1,50,1), 10)
 > My.data <- data.frame(x, y)
 > My.x <- data.frame(x=seq(5,45))
 >
 > fit <- rqss(y ~ qss(x, lambda=5), tau=0.05)
 > pred <- predict.qss1(fit, My.x)

Could someone please help me creating a dataframe "newdata" that  
would satisfy predict.qss1?

Thanks in advance,

Denis Chabot

Denis Chabot <chabotd <at> globetrotter.net> writes:

.... Quantreg package ....
> But I'd like to extract the predicted values. The help for  
> predict.qss1 indicates this:
.. 
> predict.qss1(object, newdata, ...)
> and states that newdata is a data frame describing the observations  
> at which prediction is to be made.
..
> 
>  > y <- rnorm(500, 10, 5)
>  > x <- rep(seq(1,50,1), 10)
>  > My.data <- data.frame(x, y)
>  > My.x <- data.frame(x=seq(5,45))
>  >
>  > fit <- rqss(y ~ qss(x, lambda=5), tau=0.05)
>  > pred <- predict.qss1(fit, My.x)
> 
> Could someone please help me creating a dataframe "newdata" that
> would satisfy predict.qss1?
Looks like quantreg has been reworked, but predict was not tested. 
Checking the code, I got a (hopefully) working solution

pred <- predict.qss1(fit$qss[[1]]$xyz, My.x$x)

but this follow neither convention nor documentation. Maybe you should contact
the author of the package to provide an example in the documentation, which
would have show the problem earlier.

Dieter