PreProcessor: PreProcessor

In frbl/OnlineSuperLearner: Online SuperLearner package

Description

In order to be able to use the binary / logistic loss functions, we convert every value to a value between 0 and 1. This preprocessor class takes care of this conversion by using a set of bounds (the min and max or expected min and max of the data).

Usage

PreProcessor

Format

An object of class R6ClassGenerator of length 24.

Methods

initialize(bounds)

Initializes a new PreProcessor class.

@param bounds list the bounds that should be used when normalizing the data. This list should contain an entry for each relevant variable that should be scaled. Each of those entries should then contain a min and max entry.

normalize(data)

Runs the actual normalization procedure. The data passed in is normalized according to the bounds specified on initialization.

@param data data.table the non-normalized data.

@return data.table containing the normalized data.

denormalize(data)

Runs the actual normalization procedure. The data passed in is normalized according to the bounds specified on initialization.

@param data data.table the non-normalized data.

@return data.table containing the normalized data.

frbl/OnlineSuperLearner documentation built on Feb. 9, 2020, 9:28 p.m.