defModel: Interface for defining models

In osofr/stremr: Streamlined Estimation for Static, Dynamic and Stochastic Treatment Regimes in Longitudinal Data

Description

Specify either a single model or a grid of multple models by invoking the optional argument param_grid.

Usage

defModel(estimator, x, search_criteria, param_grid, ...)

## S3 method for class 'ModelStack'
model1 + model2

Arguments

estimator	A character string name of package and estimator (algorithm) name, separated by "__".
x	A vector containing the subset of the names of the predictor variables to use in building this particular learner or grid. This argument can be used to over-ride the values of x provided to fit function. As such, the names supplied here must always be a subset of the names specified to fit. When this argument is missing (default) the column names provided to fit are used as predictors in building this model / grid.
search_criteria	Search criteria
param_grid	Named list of model hyper parameters (optional). Each named item defines either a fixed model parameter value or a vector of possible values. In the latter case this function call would define a grid (ensemble) of models. Each model in the ensemble corresponds by a particular fixed parameter value.
...	Additional modeling parameters to be passed on directly to the modeling function.
model1	An object returned by a call to defModel function.
model2	An object returned by a call to defModel function.

osofr/stremr documentation built on Jan. 25, 2022, 8:07 a.m.