Skip to main content

Table 1 Summary of the considered candidate classifiers

From: Optimal classifier selection and negative bias in error rate estimation: an empirical study on high-dimensional prediction

Method

Type

Number of genes p*

Function

Fixed parameters

Parameters tuned via CV

KNN

1

20, 50, 100, 200, 500

knnCMA

k = 1, 3, 5

 

LDA

2

10, 20

ldaCMA

  

FDA

2

10, 20

fdaCMA

  

DLDA

3

20, 50, 100, 200, 500

dldaCMA

  

PLSLDA

3

20, 50, 100, 200, 500

plsldaCMA

ncomp = 2, 3

 

NNET

3

20, 50, 100, 200, 500

nnetCMA

  

RF

4

 

rfCMA

mtry = ,, ,

 

linear SVM

4

 

svmCMA

 

cost

PAM

4

 

pamCMA

 

shrinkage parameter

L 2

4

 

plrCMA

 

penalty

  1. Column 1: Acronym of the method. Column 2: Type of the method regarding preliminary variable selection. Column 3: Number of selected genes p* (if preliminary variable selection is performed). Column 4: Name of the function in the CMA package. Column 5: Name and values of the fixed parameters. Column 6: Name of the parameters tuned using internal 3-fold-cross-validation.