EvaluationFunction: EvaluationFunction

In dGAselID: Genetic Algorithm with Incomplete Dominance for Feature Selection

Description

Evaluates the individuals' fitnesses.

Usage

EvaluationFunction(x, individuals, response, method, trainTest, nnetSize = NA,
  nnetDecay = NA, rdaAlpha = NA, rdaDelta = NA, ...)

Arguments

x	Dataset in ExpressionSet format.
individuals	Population of individuals with diploid genotypes.
response	Response variable.
method	Supervised classifier for fitness evaluation. Most of the supervised classifiers in MLInterfaces are acceptable. The default is knn.cvI(k=3, l=2).
trainTest	Cross-validation method. The default is "LOG".
nnetSize	for nnetI. The default value is NA.
nnetDecay	for nnetI. The default value is NA.
rdaAlpha	for rdaI. The default value is NA.
rdaDelta	for rdaI. The default value is NA.
...	Additional arguments.

Examples

## Not run: 
 library(genefilter)
 library(ALL)
 data(ALL)
 bALL = ALL[, substr(ALL$BT,1,1) == "B"]
 smallALL = bALL[, bALL$mol.biol %in% c("BCR/ABL", "NEG")]
 smallALL$mol.biol = factor(smallALL$mol.biol)
 smallALL$BT = factor(smallALL$BT)
 f1 <- pOverA(0.25, log2(100))
 f2 <- function(x) (IQR(x) > 0.5)
 f3 <- ttest(smallALL$mol.biol, p=0.1)
 ff <- filterfun(f1, f2, f3)
 selectedsmallALL <- genefilter(exprs(smallALL), ff)
 smallALL = smallALL[selectedsmallALL, ]
 rm(f1)
 rm(f2)
 rm(f3)
 rm(ff)
 rm(bALL)
 sum(selectedsmallALL)
 set.seed(1357)

 population0<-InitialPopulation(smallALL, 14, 8, FALSE)
 individuals0<-Individuals(population0)
 results<-EvaluationFunction(smallALL, individuals0, response="mol.biol",
             method=knn.cvI(k=3, l=2), trainTest="LOG")
 
## End(Not run)

dGAselID documentation built on May 2, 2019, 1:27 p.m.