R/targetg1.R

In benkeser/drinf: Computes doubly-robust estimators and doubly-robust confidence intervals

#' targetg1
#' 
#' Function that targets g1n.
#' 
#' @param A0 A \code{vector} treatment delivered at baseline.
#' @param A1 A \code{vector} treatment deliver after \code{L1} is measured.
#' @param L2 A \code{vector} outcome of interest. 
#' @param Qn A \code{list} of current estimates of Q2n and Q1n
#' @param gn A \code{list} of current estimates of g1n and g0n
#' @param Qnr.gnr A \code{list} of current estimates of reduced dim. regressions
#' @param abar A \code{vector} of length 2 indicating the treatment assignment 
#' that is of interest. 
#' @param tolg A \code{numeric} indicating the truncation level for conditional treatment probabilities. 
#' @param tolQ A \code{numeric}
#' @param method A character "scaled" or other
#' @param return.models  A \code{boolean} indicating whether the fluctuation model should be 
#' returned with the output.  
#' @importFrom SuperLearner trimLogit
#' 
#' @return A list with named entries corresponding to the estimators of the 
#' fluctuated nuisance parameters evaluated at the observed data values. If 
#' \code{return.models = TRUE} output also includes the fitted fluctuation model. 

targetg1 <- function(
    A0, A1, L2, Qn, gn, Qnr.gnr, 
    abar, tolg, tolQ, return.models,tol.coef=1e1, method = "scaled", ...
){
    #-------------------------------------------
    # making outcomes for logistic fluctuation
    #-------------------------------------------
    # length of output
    n <- length(gn$g0n)
    
    #-------------------------------------------
    # making offsets for logistic fluctuation
    #-------------------------------------------
    flucOff <- c(
        SuperLearner::trimLogit(gn$g1n, trim = tolg)
    )
    
    #-------------------------------------------
    # making covariates for fluctuation
    #-------------------------------------------
    # the "clever covariates" for g1n
    flucCov1 <- c(
        Qnr.gnr$Qnr$Q2nr.obsa / gn$g1n^2
    )
    #-------------------------------------------
    # making covariates for prediction
    #-------------------------------------------
    # getting the values of the clever covariates evaluated at 
    # \bar{A} = abar
    predCov1 <- c(
        Qnr.gnr$Qnr$Q2nr.seta / gn$g1n^2 # this one is set to Q2nr(abar[1], \bar{L}_1)
    )

    #-------------------------------------------
    # fitting fluctuation submodel
    #-------------------------------------------
    # first fluctuation submodel to solve original equations
    if(method != "scaled"){
        flucmod <- suppressWarnings(glm(
            formula = "out ~ -1 + offset(fo) + fc1",
            data = data.frame(out = as.numeric(A1==abar[1]), fo = flucOff, 
                              fc1 = flucCov1),
            family = binomial(), start = 0
        ))


        if(abs(flucmod$coefficients) < tol.coef){
            # get predictions 
            g1nstar <- predict(
                flucmod, 
                newdata = data.frame(out = 0, fo = flucOff,
                                     fc1 = predCov1),
                type = "response"
            )
        }else{
            g1nstar <- g1n
        }
    }else{
        # use optim to perform minimization along intercept only submodel if glm
        flucmod <- optim(
            par = 0, fn = offnegloglik, gr = gradient.offnegloglik,
            method = "L-BFGS-B", lower = -tol.coef, upper = tol.coef,
            control = list(maxit = 10000),
            Y = (as.numeric(A1==abar[2]) - tolg)/(1 - 2*tolg), 
            offset = flucOff, weight = flucCov1
        )
        epsilon <- flucmod$par
        g1nstar <- plogis(flucOff +  epsilon * flucCov1)*(1 - 2*tolg) + tolg
    }    
 
    #--------------
    # output 
    #-------------
    out <- list(
        g1nstar = g1nstar,
        flucmod = NULL
    )
    if(return.models){
        out$flucmod = list(flucmod)
    }
    return(out)
}