SMGFactory: SMGFactory

In frbl/OnlineSuperLearner: Online SuperLearner package

Description

The SMGFactory applies the Factory pattern for creating a set of SummaryMeasureGenerators, based on the needed variables. This eases the creation of summary measures as the user of this package only needs to specify the correct column names. The factory parses the formulas used and generates summary measures accordingly.

Usage

SMGFactory

Format

An object of class R6ClassGenerator of length 24.

Details

TODO: In its current form the SMG Factory is a rather static class. It is currently hard to extend this function with more SMGS without changing the actual package code. This should be improved.

Methods

initialize(candidate_smgs)

Creates a new SMG factory

@param candidate_smgs list UNUSED. This argument should be used to specify a list of possible SMGs.

fabricate(relevantVariables, ...)

Fabricates the actual SMG's. Given a list of RelevantVariables all covariates are selected and merged. These are used as a basis for selecting the needed variables in the dataframe. The column names need to be specified as follows: - just the variable name: uses the SMG.Latest.Entry class to include the contemporaneous variables. - the variable name\_lag\_the lag: uses the SMG.Lag class to include the lagged variables. - the variable name\_mean: uses the SMG.Mean class to include the running mean for this variable.

@param relevantVariables = a list of RelevantVariable objects, from which the X variables are selected.

@param ... = data to get passed to the SMG

fabricate(relevantVariables, ...)

Fabricates the actual SMG's. Given a list of RelevantVariables all covariates are selected and merged. These are used as a basis for selecting the needed variables in the dataframe.

@param relevantVariables list a list of RelevantVariable objects, from which the X variables are selected.

@param ... the data to get passed to the SMG.

@return SummaryMeasureGenerator with the correct SMGs.

get_candidate_smgs

Active method. Returns the candidate SummaryMeasureGenerators provided on initialization.

@return list the list of candidate_smgs.

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