mcEvalPsi: R6 class for Monte-Carlo evaluation of various substitution...
In tmlenet: Targeted Maximum Likelihood Estimation for Network Data

Description Usage Format Details Methods

This R6 class performs the Monte-Carlo evaluation of the target parameters using the data generated under the user-specified arbitrary intervention gstar. For a given dataset, take E[Y|sA,sW] = m.Q.init and calcualte estimate of psi_n under g_star using Monte-Carlo integration: (*) W can be iid or not (W's are never resampled). (*) Use P_n(W) = 1 for the distribution of W = (W_1,...,W_n) and draw n new exposures A=(A_1,...,A_n) from the distribution of g_star. (*) Evaluate n summary measures sA=(sA_1,...,sA_n) using these n newly sampled exposures A. (*) Evaluate the subsititution estimators as an average of n predictions E[Y=1|sA,sW]. (*) Repeat nrep times until convergence of psi_n.