el.est: Empirical Likelihood-Based Estimates

In kimihua1995/adjKMtest:

Description

Estimate survival probabilities by empirical likelihood-based (EL) method

Usage

el.est(y,delta,treat,x,psix_moment = c("first","second"),
treat.select,t,get.sd=TRUE,Nboot=500,standardize=FALSE)

Arguments

y	a numeric vector of observed time
delta	a factor vector of indicator for events
treat	a factor vector of the treatment indicator
x	an optional covariance matrix containing the variables in the model
psix_moment	first or second, which correspond to adjusting first or second moments of the covariates matrix x
treat.select	a value of 1 (treatment) or 0 (control)
t	a numeric vector of time points for estimation
get.sd	a logical object of whether return bootstrap standard deviation, default by TRUE
Nboot	number of iterations in Bootstrap, default by 500
standardize	a logical object of whether the covariance matrix x would be centered and scaled, default by FALSE

Value

A numeric vector with the same length of treat, which contains the propensity scores for each observations.

Examples

library(survival)
y <- as.matrix(gbsg$rfstime/(365.25/12))
treat <- data.matrix(gbsg$hormon)
delta <- as.matrix(gbsg$status)
x <- data.matrix(gbsg[,c("age","meno","size","grade","nodes","pgr","er")])
t <- seq(0,80,2)

el.est(y=y,delta=delta,treat=treat,x=x,psix_moment="first",treat.select=1,t=t,
                   get.sd=TRUE,Nboot=500,standardize=FALSE)

kimihua1995/adjKMtest documentation built on Jan. 1, 2021, 7:20 a.m.