outer.convhull(cx,cy,px,py,FUN,duplicate="remove",...)
cx
| x cordinates of grid |
cy
| y cordinates of grid |
px
| vector of x coordinates of points |
py
| vector of y coordinates of points |
FUN
| function to be evaluated over the grid |
duplicate
|
indicates what to do with duplicate
(px_i,py_i) points, default "remove".
|
outer evaluates FUN
only on that part of the grid cx x cy that is enclosed within
the convex hull of the points (px,py).
This can be useful for spatial estimation if no extrapolation is wanted.
FUN (NAs if outside the
convex hull).in.convex.hull
x<-runif(20)
y<-runif(20)
z<-runif(20)
z.lm<-lm(z~x+y)
f.pred<-function(x,y)
{predict(z.lm,data.frame(x=as.vector(x),y=as.vector(y)))$predictor}
xg<-seq(0,1,0.05)
yg<-seq(0,1,0.05)
image(xg,yg,outer.convhull(xg,yg,x,y,f.pred))
points(x,y)