Description Usage Format Details Methods See Also
MonteCarloSimClass
only resamples A
under the intervention function g_star
, never W
or E
.
For each MC simulation, it firstly treats model.Q.init
as the fitted models for E[Y|A,W,E]
, then estimate
psi_n
using Monte-Carlo integration. i.e., average of n
predicted E(Y|A=a, W=w,E=e)
where a
is
a vector of n
new exposures randomly drawn under g_star
. Take as many iterations as needed until convergence
of ψ^{I}_n (or ψ^{II}_n) occurs.
1 |
An R6Class
generator object
OData.ObsP0
- A DatKeepClass
class object, where exposures are generated under observed exposure mechanism g0.
OData.gstar
- A DatKeepClass
class object, where new exposures are generated under user-specified
intervention g^{*}.
model.Q.init
- A fitted model for E[Y|A,W,E]
.
model.Q.star.cov
- A targeting model for covariate-based unweighted TMLE.
model.Q.star.int
- A targeting model for intercept-based TMLE.
QY.init
- Estimates of G-COMP.
QY.star.cov
- Estimates of covariate-based unweighted TMLE.
QY.star.int
- Estimates of intercept-based TMLE.
nobs
- Number of observations in the observed data frame.
p
- Number of Monte-Carlo simulations performed.
new(OData.ObsP0, OData.gstar, ...)
Instantiate an new instance of MonteCarloSimClass
.
get.gcomp(m.Q.init)
Predict QY.init
= E[Y_{g^*}] based on the initial model fit model.Q.init
.
get.tmleCov(model.Q.star.cov, model.h.fit)
Update QY.init
based on the targeting model model.Q.star.cov
and the model for clever covriate h model.h.fit
.
get.tmleCov(model.Q.star.cov, model.h.fit)
Update QY.init
based on the targeting model model.Q.star.cov
and the model for clever covriate h model.h.fit
.
get.tmleInt(model.Q.star.int)
Update QY.init
based on the targeting model model.Q.star.int
.
get.fiW()
Get an estimate of fiW (hold ALL W's fixed).
tmleCom_Options
, DatKeepClass
, RegressionClass
, tmleCommunity
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.