Dear All,
is there a way to superimpose points and/or lines on a surface plot?
Below I try to explain my problem.
Suppose I have the following surface plot (likelikood for the normal
variable when both parameters are unknown):
-------------------------------------------------------------------------------------------------------------------------------------------
normalLike <- function(mu, sigma, sample){
(sigma ^ - length(sample)) * exp(-0.5 * (sigma ^ -2) * sum((sample -
mu)^2))
}
normalLikeVec <- Vectorize(normalLike, vectorize.args =
c("mu","sigma"))
teta1 <- seq(0, 6, by=0.1)
teta2 <- seq(0.1, 8, by=0.1)
matrixMuSigma <- outer(teta1, teta2, normalLikeVec, sample=c(0.88, 1.07,
1.27,1.54, 1.91, 2.27, 3.84, 4.50, 4.64, 9.41))
matrixMuSigma <- matrixMuSigma / max(matrixMuSigma)
persp(teta1, teta2, matrixMuSigma)
-------------------------------------------------------------------------------------------------------------------------------------------
Then I would like to superimpose on the same plot points (or a line)
corresponding to the following data:
-------------------------------------------------------------------------------------------------------------------------------------------
#x coordinates
teta1
#y coordinates
sigmaMax <- teta2[apply(matrixMuSigma, 1, which.max)]
#z coordinates
apply(matrixMuSigma, 1, max)
-------------------------------------------------------------------------------------------------------------------------------------------
Thanks in advance,
domenico
PS:
I am trying to explain the geometrical interpretation of profile
likelihood. Any better idea?
See ?persp, particularly the example that starts with # (2) Add to existing persp plot - using trans3d() : Uwe Ligges Domenico Vistocco wrote:> Dear All, > is there a way to superimpose points and/or lines on a surface plot? > > Below I try to explain my problem. > Suppose I have the following surface plot (likelikood for the normal > variable when both parameters are unknown): > > ------------------------------------------------------------------------------------------------------------------------------------------- > > normalLike <- function(mu, sigma, sample){ > (sigma ^ - length(sample)) * exp(-0.5 * (sigma ^ -2) * sum((sample - > mu)^2)) > } > normalLikeVec <- Vectorize(normalLike, vectorize.args = c("mu","sigma")) > > teta1 <- seq(0, 6, by=0.1) > teta2 <- seq(0.1, 8, by=0.1) > matrixMuSigma <- outer(teta1, teta2, normalLikeVec, sample=c(0.88, 1.07, > 1.27,1.54, 1.91, 2.27, 3.84, 4.50, 4.64, 9.41)) > matrixMuSigma <- matrixMuSigma / max(matrixMuSigma) > > persp(teta1, teta2, matrixMuSigma) > ------------------------------------------------------------------------------------------------------------------------------------------- > > > Then I would like to superimpose on the same plot points (or a line) > corresponding to the following data: > ------------------------------------------------------------------------------------------------------------------------------------------- > > #x coordinates > teta1 > #y coordinates > sigmaMax <- teta2[apply(matrixMuSigma, 1, which.max)] > #z coordinates > apply(matrixMuSigma, 1, max) > ------------------------------------------------------------------------------------------------------------------------------------------- > > > Thanks in advance, > domenico > > PS: > I am trying to explain the geometrical interpretation of profile > likelihood. Any better idea? > > ______________________________________________ > R-help at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide > http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.
In this case, following your instructions and a bit of noodling with the example "solves" the problem with: persp(teta1, teta2, matrixMuSigma) -> res points(trans3d( teta1, sigmaMax, apply(matrixMuSigma, 1, max), pmat=res), col="red") -- David Winsemius On Feb 10, 2009, at 11:50 AM, Uwe Ligges wrote:> See ?persp, particularly the example that starts with > # (2) Add to existing persp plot - using trans3d() : > > Uwe Ligges > > > Domenico Vistocco wrote: >> Dear All, >> is there a way to superimpose points and/or lines on a surface plot? >> Below I try to explain my problem. >> Suppose I have the following surface plot (likelikood for the >> normal variable when both parameters are unknown): >> ------------------------------------------------------------------------------------------------------------------------------------------- normalLike >> <- function(mu, sigma, sample){ >> (sigma ^ - length(sample)) * exp(-0.5 * (sigma ^ -2) * >> sum((sample - mu)^2)) >> } >> normalLikeVec <- Vectorize(normalLike, vectorize.args = >> c("mu","sigma")) >> teta1 <- seq(0, 6, by=0.1) >> teta2 <- seq(0.1, 8, by=0.1) >> matrixMuSigma <- outer(teta1, teta2, normalLikeVec, sample=c(0.88, >> 1.07, 1.27,1.54, 1.91, 2.27, 3.84, 4.50, 4.64, 9.41)) >> matrixMuSigma <- matrixMuSigma / max(matrixMuSigma) >> persp(teta1, teta2, matrixMuSigma) >> ------------------------------------------------------------------------------------------------------------------------------------------- Then >> I would like to superimpose on the same plot points (or a line) >> corresponding to the following data: >> ------------------------------------------------------------------------------------------------------------------------------------------- #x >> coordinates >> teta1 >> #y coordinates >> sigmaMax <- teta2[apply(matrixMuSigma, 1, which.max)] >> #z coordinates >> apply(matrixMuSigma, 1, max) >> ------------------------------------------------------------------------------------------------------------------------------------------- Thanks >> in advance, >> domenico >> PS: >> I am trying to explain the geometrical interpretation of profile >> likelihood. Any better idea? >> ______________________________________________ >> R-help at r-project.org mailing list >> https://stat.ethz.ch/mailman/listinfo/r-help >> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html >> and provide commented, minimal, self-contained, reproducible code. > > ______________________________________________ > R-help at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.