R help - May 2005 - persp( ) Question

I have successfully fitted the model

    loess.fit1 <- loess(response ~ X*Y)

and plotted it in 3D using


X.grid <- seq(0,10,length=100)
Y.grid <- seq(0,1000,length=100)
pred.loess1 <- predict(loess.fit1, 
        expand.grid(x = X.grid, y = Y.grid))
persp(X.grid, Y.grid, pred.loess1, theta = 0, phi 12)


I would like to add a series of points along the
fitted surface at X.grid = 2, in red.  The example
in the online help uses


f <- function(x,y) 
  { r <- sqrt(x^2+y^2); 10 * sin(r)/r }
trans3d <- function(x,y,z, pmat) 
{
  tr <- cbind(x,y,z,1) %*% pmat
  list(x = tr[,1]/tr[,4], y= tr[,2]/tr[,4])
}
phi <- seq(0, 2*pi, len = 201)
r1 <- 7.725 # radius of 2nd maximum
xr <- r1 * cos(phi)
yr <- r1 * sin(phi)
lines(trans3d(xr,yr, f(xr,yr), res), 
      col = "pink", lwd=2)


to add points to the fitted mathematical surface. I
have not been able to reproduce this sort of 
functionality partly because I don't have a functional
representation of my surface and partly because I
don't understand what trans3d( ) is doing.  Any help
would be greatly appreciated.


Kind regards,

    Greg




		

Stay connected, organized, and protected. Take the tour:

Greg,

Assign the output of "persp" to a variable "pmat":
R> pmat <- persp(X.grid, Y.grid, pred.loess1, theta=0, phi=12)

Now you can add points to your plot with the usual "points" command.
But you have to translate your 3D coordinates (x,y,z) into 2D
coordinates for "points" to understand, and that's what
"trans3d" does:

R> points(trans3d(x,y,z, pmat), col="red")

You supply the (x,y,z) values, of course.  It's a mystery to me why
"trans3d" is not included in the graphics package (where
"persp" lives).
HTH.

-- David Brahm (brahm at alum.mit.edu)