QlearnModel: R6 Class for Q-Learning

In stremr: Streamlined Estimation of Survival for Static, Dynamic and Stochastic Treatment and Monitoring Regimes

Description

R6 class for controlling the internal implementation of Q-learning functionality. Supports sequential (recursive) G-computation and longitudinal TMLE. Inherits from BinaryOutcomeModel R6 Class.

Usage

QlearnModel

Format

An R6Class generator object

Details

• regimen_names - character. Note used. For future pooling across regimens.

• classify - ... logical

• TMLE - ... logical

• nIDs - ... integer

• stratifyQ_by_rule - ... logical

• lower_bound_zero_Q - ... logical

• skip_update_zero_Q - ... logical

• Qreg_counter - ... integer

• t_period - ... integer

• idx_used_to_fit_initQ - ... integer

Methods

new(reg, ...)

...

define.subset.idx(data, subset_vars, subset_exprs)

...

define_idx_to_fit_initQ(data)

...

define_idx_to_predictQ(data)

...

fit(overwrite = FALSE, data, ...)

...

Propagate_TMLE_fit(overwrite = TRUE, data, new.TMLE.fit, ...)

...

predict(newdata, subset_idx, ...)

...

predictStatic(data, g0, gstar, subset_idx)

...

predictStochastic(data, g0, gstar, subset_idx, stoch_indicator)

...

predictAeqa(newdata, ...)

...

get.fits()

...

Active Bindings

wipe.alldat

...

getfit

...

getTMLEfit

...

stremr documentation built on May 30, 2017, 6:35 a.m.