[R] 2D contour predictions

[Michael Hopkins]

Hi All

I have been fitting regression models and would now like to produce some
contour & image plots from the predictors.

Is there an easy way to do this?  My current (newbie) experience with R
would suggest there is but that it's not always easy to find it!

f3 <- lm( fc ~ poly( speed, 2 ) + poly( torque, 2 ) + poly( sonl, 2 ) +
poly( p_rail, 2 ) + poly( pil_sep, 2 ) + poly( maf, 2 ) + (speed + torque +
sonl + p_rail + pil_sep + maf)^2 )

hat <- predict( f3 )

contour( sonl, maf, hat )

Error in contour.default(sonl, maf, hat) :
    increasing 'x' and 'y' values expected

image(sonl, maf, hat)

Error in image.default(sonl, maf, hat) : increasing 'x' and 'y' values
expected

I have tried naïve sorting but no luck with that.

I suspect I may need to produce a data grid of some kind but I'm not clear
how I would use R to specify such a 2D slice in the 6D design space.

TIA

Michael

P.S.  Whilst I'm asking the list - is there an easier way of expressing a
full 2nd order model than the one I used above?  I'm sure there is but
finding it etc etc...


_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/

        _/    _/   _/_/_/             Hopkins Research Ltd
       _/    _/   _/    _/
      _/_/_/_/   _/_/_/          http://www.hopkins-research.com/
     _/    _/   _/   _/
    _/    _/   _/     _/               'touch the future'
                   
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/