classify: Wrapper to run all classifiers for one run

In nlcv: Nested Loop Cross Validation

Description

Wrapper function that will run all classification algorithms using a given split in training and test data (corresponding to one run in the outer loop of the double cross validation).

Usage

classify(eset, trainingSample, testSample, classVar = "type")

Arguments

eset	Expression set on which the classifiers will be run
trainingSample	matrix with for each run the indices of the observations included in the training (or learning) sample
testSample	matrix with for each run the indices of the observations included in the test (or validation) sample
classVar	String giving the name of the variable containing the observed class labels

Value

A list with the following components

dlda	estimated misclassification rate using diagonal linear discriminant analysis
svm	estimated misclassification rate using support vector machines
randomForest	estimated misclassification rate using a random forest
bagg	estimated misclassification rate using bagging
pam	estimated misclassification rate using the pam algorithm
dlda.predic	predicted values by linear discriminant analysis
svm.predic	predicted values by support vector machines
randomForest.predic	predicted values by random forests
bagg.predic	predicted values by bagg
pam.predic	predicted values by the pam algorithm

Note

The different classification algorithms are called via the uniform interfaces of the MLInterfaces package.

Author(s)

Willem Talloen and Tobias Verbeke

nlcv documentation built on May 2, 2019, 4:56 p.m.