Regression diagnostics for trend surfaces

DESCRIPTION:

This function provides the basic quantities which are used in forming a variety of diagnostics for checking the quality of regression fits for trend surfaces calculated by surf.ls.

USAGE:

trls.influence(object)
plot.trls(x, border = 1, col = NA, pch = 4, cex = 0.6,
          add = F, div = 8, ...)

REQUIRED ARGUMENTS:

object, x
Fitted trend surface model from surf.ls

OPTIONAL ARGUMENTS:

div
scaling factor for influence circle radii in plot.trls
add
add influence plot to existing graphics if TRUE
border, col, pch, cex, ...
additional graphical parameters

VALUE:

trls.influence returns a list with:
r
raw residuals as given by residuals.trls
hii
diagonal elements of the Hat matrix
stresid
standardised residuals
Di
Cook's statistic

REFERENCES:

Unwin, D. J., Wrigley, N. (1987) Towards a general-theory of control point distribution effects in trend surface models. Computers and Geosciences, 13, 351-355.

SEE ALSO:

, ,

EXAMPLES:

library(MASS)
topo2 <- surf.ls(2, topo)
infl.topo2 <- trls.influence(topo2)
cand <- as.data.frame(infl.topo2)[abs(infl.topo2$stresid) > 1.5,]
cand
cand.xy <- topo[as.integer(dimnames(cand)[[1]]), c("x", "y")]
trsurf <- trmat(topo2, 0, 6.5, 0, 6.5, 50)
eqscplot(trsurf, type="n")
#under S need to choose appropriate colour numbers
contour(trsurf, add=T, col="grey")
plot(topo2, add=T, div=3)
points(cand.xy, pch=16, col="orange")