arcfsReweighter: Reweighter for the arc-fs algorithm.

In boostr: A modular framework to bag or boost any estimation procedure.

Description

A slightly modified implementation of the reweighter described in the arc-fs algorithm.

Usage

arcfsReweighter(prediction, response, weights, ...)

Arguments

prediction	a vector of predictions.
response	a vector whose ith component is the true response for the ith component of prediction.
weights	a vector of weights. They don't necessarily need to sum to 1.
...	implemented to allow reweighter to accept its output as its input.

Details

As per Leo Breiman's suggestions, a slight modification to the arc-fs algorithm has been made in the case where ε – the misclassification measure – exceed 0.5, or becomes 0. Should this happen β is set to -∞ and a warning is produced. At this point you are advised to restart the algorithm with equal probabilities or stop boosting, at that iteration.

Value

For internal bookkeeping, this function is inherits from the 'reweighter' class. It returns a named list with components

weights	the updated weights calculated from the input weights, prediction and response.
beta	scalar weights to be used by arcfsAggregator.

See Also

Other arc-fs: arcfsAggregator

Other reweighters: adaboostReweighter; arcx4Reweighter; boost, boost.function, boost.list; vanillaBagger

boostr documentation built on May 2, 2019, 1:42 p.m.