# Estimate Quantile-optimal Treatment Regime for covariates-dependent random censoring data This function estimates the Quantile-optimal Treatment Regime for a given quantile level of interest under the assumption that the distribution of censoring time is independent of the set of potential survival times given a set of baseline covariates and treatment actually received. More specifically, we do stratification by treatment first and then used kernel smoothing to estimate local survival function of censoring time for each treatment group. ```r IPWE_Qopt_DepCen_general(data, regimeClass, tau, Domains = NULL, bw, moPropen = "BinaryRandom", DepCens = NULL, UseTrueG = FALSE, trueG_value = NULL, cluster = FALSE, p_level = 1, s.tol = 1e-05, it.num = 8, pop.size = 5000) ``` ## Arguments - `data`: raw data.frame - `regimeClass`: the class of treatment regimes. e.g., 'txname ~ x1+x2'. - `tau`: the quantile of interest - `Domains`: default is NULL. - `bw`: the bandwidth of local KM model (e.g. see Wang-wang 2008) - `moPropen`: an optional string for the working model of treatment assignment - `DepCens`: an optional vector of baseline variable names that the censoring variable depends on. Note that the treatment variable is always treated as dependent with the censoring time. If unspecified (`DepCens=NULL`), then all variables on the right side of `regimeClass` are used for `DepCens` - `UseTrueG`: logical. Whether the true survival probability of each patient is provided. - `trueG_value`: default is NULL. IF `UseTrueG=FALSE`, `trueG_value` should be `NULL`. - `cluster`: default is FALSE. This can also be an object of the 'cluster' class returned by one of the makeCluster commands in the parallel package or a vector of machine names so rgenoud::genoud can setup the cluster automatically. - `p_level`: print level - `s.tol`: tolerance level - `it.num`: the maximum iteration number - `pop.size`: the initial population size ## Examples ```r GenerateData <- function(n) { x1 <- runif(n, min=-0.5,max=0.5) x2 <- runif(n, min=-0.5,max=0.5) error <- rnorm(n, sd= 1) ph <- exp(-0.5+1*(x1+x2))/(1+exp(-0.5 + 1*(x1+x2))) a <- rbinom(n = n, size = 1, prob=ph) c <- 1.5 + + runif(n = n, min=0, max=2) cmplt_y <- pmin(2+x1+x2 + a*(1 - x1 - x2) + (0.2 + a*(1+x1+x2)) * error, 4.4) censor_y <- pmin(cmplt_y, c) delta <- as.numeric(c > cmplt_y) return(data.frame(x1=x1,x2=x2,a=a, censor_y = censor_y, delta=delta)) } n <- 400 data <- GenerateData(n) fit1 <- IPWE_Qopt_DepCen_general(data = data, regimeClass = a~x1+x2, moPropen = a~x1+x2, tau = 0.2, bw = 20/n, pop.size=3000, it.num = 3) ```

IPWE_Qopt_DepCen_general function