Rmagpie (version 1.28.0)

thresholds-class: thresholds: A class to handle the thresholds to be tested during training of the Nearest Shrunken Centroid

Description

The Nearest Shrunken Centroid is computed using a threshold. This threshold is usually determined by finding the best threshold value over a set of values by finding the threshold leading to the best error rate assessed by cross-validation. This class stores the values of thresholds to be tried. If the user wants to use default values it's also possible.

Creating objects

new("thresholds") Create an empty thresholds. The default thresholds values will be computed and this object updated as soon as it is linked in an assessment. new("thresholds", optionValues) Create a thresholds, containing the thresholds values defined by optionValues. The slot noOfOptions is automatically updated.

Slots

optionValues:

numeric Values of the thresholds, if optionValues has length zero then the default thresholds values must be used.

noOfOptions:

numeric Number of thresholds.

Extends

Class "featureSelectionOptions", directly.

Methods

getNoThresholds(thresholds)

Retreive the number of the thresholds (slot noOfOptions)

getOptionValues(thresholds), getOptionValues(thresholds)<-

Retreive and modify the values of the thresholds (slot optionValues)

See Also

geneSubsets, assessment

Examples

# Empty thresholds, the default values will be used when added to an assessment
emptThresholds <- new("thresholds")
getOptionValues(emptThresholds)
getNoThresholds(emptThresholds)

# Another thresholds
thresholds <- new("thresholds", optionValues=c(0,0.1,0.2,1,2))
getOptionValues(thresholds)
getNoThresholds(thresholds)

# Set the thresholds
newThresholds <- c(0.1,0.2,0.5,0.6,1)
getOptionValues(thresholds) <- newThresholds
getOptionValues(thresholds)
getNoThresholds(thresholds)