library("mlt") library("survival") options(digits = 3) data("GBSG2", package = "TH.data") xvar <- names(GBSG2) xvar <- xvar[!(xvar %in% c("time", "cens"))] GBSG2$y <- with(GBSG2, Surv(time, cens)) fm <- as.formula(paste("Surv(time, cens) ~ ", paste(xvar, collapse = "+"))) cmod <- coxph(fm, data = GBSG2) order <- 10 by <- Bernstein_basis(numeric_var("y", support = c(0, max(GBSG2$time))), order = order, ui = "incre") bx <- as.basis(as.formula(paste("~", paste(xvar, collapse = "+"))), data = GBSG2, remove_intercept = TRUE) m <- ctm(by, shift = bx, todist = "MinEx") mod <- mlt(m, data = GBSG2, scale = TRUE) n <- names(coef(cmod)) cf <- coef(mod)[n] v <- vcov(mod)[n, n] coef(cmod) / cf diag(vcov(cmod)) / diag(v) range(vcov(cmod) / v)