important.variable.names: Writing tables with variable information

In MCRestimate: Misclassification error estimation with cross-validation

Description

The information slot of a MCRestimate object may contain lists of variable names that are important for each classification. This function produces summary tables of these variables. It is assumed that the first column of each information list contains the variable names if not otherwise specified with the argument listName. This is important if someone wants to write a new wrapper for a classification method.

Usage

important.variable.names(mcr,file="important_variables",listName=NULL,writeFile=TRUE,...)

Arguments

mcr	an object of class MCRestimate
file	a character string specifies the name of the output files
listName	a character string specifying the variable names vector in the information list
writeFile	Should the files be written?
...	Further arguments that are passed on to plot.default

Value

The function is called for its side effect, writing two tables with variable information.

Author(s)

Markus Ruschhaupt mailto:[e-mail]

Examples

library(golubEsets)
data(Golub_Train)


class.column <- "ALL.AML"
list.of.poss.parameter <- list(var.numbers = c(250,1000))
Preprocessingfunctions <- c("identity")
list.of.poss.parameter <- list(threshold = 6)
class.function <- "PAM.wrap"
plot.label <- "Samples"

cross.outer <- 10
cross.repeat <- 7
cross.inner <- 5

PAM.estimate <- MCRestimate(Golub_Train,
		class.column,
		classification.fun = class.function,
		thePreprocessingMethods = Preprocessingfunctions,
		poss.parameters = list.of.poss.parameter,
		cross.outer = cross.outer,
		cross.inner = cross.inner,
		cross.repeat = cross.repeat,
		plot.label = plot.label)

		important.variable.names(PAM.estimate)

MCRestimate documentation built on Oct. 31, 2019, 10:29 a.m.