mlbench (version 2.1-3)

bayesclass: Bayes classifier

Description

Returns the decision of the (optimal) Bayes classifier for a given data set. This is a generic function, i.e., there are different methods for the various mlbench problems.

If the classes of the problem do not overlap, then the Bayes decision is identical to the true classification, which is implemented as the dummy function bayesclass.noerr (which simply returns z$classes and is used for all problems with disjunct classes).

Usage

bayesclass(z)

Arguments

z

An object of class "mlbench".

Examples

# NOT RUN {
# 6 overlapping classes
p <- mlbench.2dnormals(500,6)
plot(p)

plot(p$x, col=as.numeric(bayesclass(p)))
# }