plotModel: Surface plot of a model

In bartzbeielstein/SPOT: Sequential Parameter Optimization Toolbox

Description

A (filled) contour or perspective plot of a fitted model.

Usage

plotModel(
  object,
  which = if (ncol(object$x) > 1 & tolower(type) != "singledim") {     1:2 } else {    
    1 },
  constant = object$x[which.min(object$y), ],
  xlab = paste("x", which, sep = ""),
  ylab = "y",
  type = "filled.contour",
  ...
)

Arguments

object	fit created by a modeling function, e.g., buildRandomForest.
which	a vector with two elements, each an integer giving the two independent variables of the plot (the integers are indices of the respective data set).
constant	a numeric vector that states for each variable a constant value that it will take on if it is not varied in the plot. This affects the parameters not selected by the which parameter. By default, this will be fixed to the best known solution, i.e., the one with minimal y-value, according to which.min(object$y). The length of this numeric vector should be the same as the number of columns in object$x
xlab	a vector of characters, giving the labels for each of the two independent variables.
ylab	character, the value of the dependent variable predicted by the corresponding model.
type	string describing the type of the plot: "filled.contour" (default), "contour", "persp" (perspective), or "persp3d" plot. Note that "persp3d" is based on the plotly package and will work in RStudio, but not in the standard RGui.
...	additional parameters passed to the contour or filled.contour function.

See Also

plotFunction, plotData

Examples

## generate random test data
testfun <- function (x) sum(x^2)
set.seed(1)
k <- 30
x <- cbind(runif(k)*15-5,runif(k)*15,runif(k)*2-7,runif(k)*5+22)
y <- as.matrix(apply(x,1,testfun))
fit <- buildLM(x,y)
plotModel(fit)
plotModel(fit,type="contour")
plotModel(fit,type="persp")
plotModel(fit,which=c(1,4))
plotModel(fit,which=2:3)

bartzbeielstein/SPOT documentation built on June 13, 2020, 5:58 p.m.