CovComp: Output the martingale and counting process covariance for two...

In blangworthy/survPCA: Estimation of PCA for Survival Data

Description

This takes survival times and censoring/cause indicators for two variables and outputs the martingale and counting process covariance at time t

Usage

CovComp(data, t, cause1, cause2, bivsurvtype = "dabrowska")

Arguments

data	An n by 4 matrix where the first column is the observed event time of the first variable, the second column is the observed event time of the second variable, the third column is the censoring/cause indicator for the first variable (0 indicates censored) and the fourth column is the censoring/cause indicator for the second variable (0 indicates censored)
t	The time to evaluate the martingale and counting process covariance at
cause1	An indicator of which cause to calculate the cause specific hazard for for the first variable, should be a non-zero value that appears in the censoring/cause indicator column for the first variable
cause2	An indicator of which cause to calculate the cause specific hazard for for the second variable, should be a non-zero value that appears in the censoring/cause indicator column for the second variable
bivsurvtype	A string vector indicating the type of bivariate survival estimator to use, "dabrowska" for Dabrowska's estimator and "linying" for Lin Ying. Lin Ying should only be used for univariate censoring, Dabrowska is default

Value

A list with the following elements

• MartCov: The martingale covariance between variable 1 and variable 2 at time t

• CountCov: The counting process covariance between variable 1 and variable 2 at timet

• CIFVar1: The cause specifice cumulative incidence function estimate for variable 1

• Survar2: The cause specifice cumulative incidence function estimate for variable 2

blangworthy/survPCA documentation built on Oct. 3, 2020, 3:04 p.m.