valipars: Generate Control Parameters For Validation / Resampling

Description Usage Arguments Details Value Author(s) See Also Examples

Description

Generate the control parameters for resampling or validation process.

Usage

1
valipars(sampling="cv", niter=10, nreps=10, strat=FALSE,div = 2/3) 

Arguments

sampling

Sampling scheme. Valid options are:

  • loocv. Leave-one-out cross-validation

  • cv. K-fold cross-validation (default)

  • rand. Randomised validation (holdout)

  • boot. Bootstrap

niter

Number of iteration or repeat for validation.

nreps

Number of replications in each iteration (number of folds for sampling=cv and bootstrap for sampling=boot).

strat

A logical value indicating if stratification is applied to sampling=cv, rand and boot.

div

Proportion of training data randomly selected for sampling=rand.

Details

valipars provides a list of control parameters for the resampling or validation in the process of accuracy evaluation or feature selection process.

Value

An object of class valipars containing all the above parameters (either the defaults or the user specified values).

Author(s)

Wanchang Lin [email protected]

See Also

trainind, accest

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
## generate control parameters for the re-sampling scheme with 5-fold 
## cross-validation and iteration of 10 times
valipars(sampling = "cv", niter = 10, nreps = 5)

## generate control parameters for the re-sampling scheme with 
## 25-replication bootstrap and iteration of 100 times
valipars(sampling = "boot", niter = 100, nreps = 25,strat=TRUE)

## generate control parameters for the re-sampling scheme with 
## leave-one-out cross-validation
valipars(sampling = "loocv")

wilsontom/FIEmspro documentation built on Feb. 19, 2018, 9:03 a.m.