sindex: Index for evaluation of step functions.

In prodlim: Product-Limit Estimation for Censored Event History Analysis

Index for evaluation of step functions.

Description

Returns an index of positions. Intended for evaluating a step function at selected times. The function counts how many elements of a vector, e.g. the jump times of the step function, are smaller or equal to the elements in a second vector, e.g. the times where the step function should be evaluated.

Usage

sindex(jump.times, eval.times, comp = "smaller", strict = FALSE)

Arguments

jump.times	Numeric vector: e.g. the unique jump times of a step function.
eval.times	Numeric vector: e.g. the times where the step function should be evaluated
comp	If "greater" count the number of jump times that are greater (greater or equal when strict==FALSE) than the eval times
strict	If TRUE make the comparison of jump times and eval times strict

Details

If all jump.times are greater than a particular eval.time the sindex returns 0. This must be considered when sindex is used for subsetting, see the Examples below.

Value

Index of the same length as eval.times containing the numbers of the jump.times that are smaller than or equal to eval.times.

Author(s)

Thomas A. Gerds [e-mail]

Examples

test <- list(time = c(1, 1,5,5,2,7,9),
		status = c(1,0,1,0,1,1,0))
fit <- prodlim(Hist(time,status)~1,data=test)
jtimes <- fit$time
etimes <- c(0,.5,2,8,10)
fit$surv
c(1,fit$surv)[1+sindex(jtimes,etimes)]

prodlim documentation built on June 24, 2024, 5:08 p.m.