loo: Leave-one-out cross-validation

Description Usage Arguments Details Value Examples

View source: R/loo.R

Description

Fitting and validation of a predictor in a leave-one-out protocol.

Usage

1
  loo(eset, class="class", method = "welch.test", ngenes=50, dist="cor", hparam = 0.75, positive="") 

Arguments

eset

Bioconductor ExpressionSet

class

String specifying the column in pData(eset) that contains the class information.

method

Specifying the feature selection method. Possible values are "cor", "student.test", "welch.test", "wilcoxon.test", "foldchange", "copa", "os", "ort", "shift", "throw".

ngenes

Number of features used for classification.

dist

Metric for distance calculation

hparam

Hyperparameter needed for some of the feature selection methods. For methods copa, ors and os: Quantile (e.g. 0.75, 0.9, 0.95) used in order to detect outliers. For methods shift and throw: the minimum number of samples in each class after applying shift or throw.

positive

One of the two classes. Membership to this class is considered as positive. Needed in order to calculate sensitivity and specificity of the validation.

Details

A leave-one-out cross-validation is performend by calling fit and predict in a loop.

Value

A pvalidation object, see pvalidation.object for details.

Examples

1
### see: help(GOLUB);

Bioconductor-mirror/cancerclass documentation built on June 1, 2017, 5:25 a.m.