Question

I need to mark where certain observations appear in a 3D plotted joint density function — I envision adding a vector, (x, y, f(x,y) + something_small) to the density plot, showing where the point is. I have tried using trans3d(), but that hasn’t worked.

Here is an example:

library(MASS)
Sigma <- matrix(c(12,1,1,12),2,2)
Sample <- mvrnorm(n=1000, rep(0, 2), Sigma)
empDen <- kde2d(Sample[,1],Sample[,2])
par(bg = "white")

x <- empDen$x
y <- empDen$y
z <- empDen$z
nrz <- nrow(z)
ncz <- ncol(z)
jet.colors <- colorRampPalette( c("lightblue", "blue") ) 
nbcol <- 100
color <- jet.colors(nbcol)
zfacet <- z[-1, -1] + z[-1, -ncz] + z[-nrz, -1] + z[-nrz, -ncz]
facetcol <- cut(zfacet, nbcol)
persp(x, y, z, col = color[facetcol], phi = 15, theta = -50, xlab="x", ylab="y", zlab="Empirical Joint Density", border=NA)

The question is: How do I indicate where Sample[1,] appears in the joint density, i.e. add this to the plot?

Thanks for any tips!

Answer 1

This works:

fmt=persp(x, y, z, col = color[facetcol], phi = 15, theta = -50, xlab="x", ylab="y", zlab="Empirical Joint Density", border=NA)
pt = Sample[1,]
points(trans3d(pt[1],pt[2],.001,fmt),pch=20, col="Red")
lines(trans3d(c(pt[1],pt[1]), c(pt[2],pt[2]), c(0,.001),fmt),col="Red",cex=2) 

Although, it would be nice to replace .001 with some information based off the empirical joint density instead of manually specifying values for each point.