survCompeting: Self-Consistent, Competing Risks (SC-CR) Algorithms

In sccr: The Self-Consistent, Competing Risks (SC-CR) Algorithms

Description

This package describes an algorithm for producing fully non-parametric and self-consistent estimators of the cause-specific failure probabilities in the presence of interval-censoring and possible masking of the failure cause in a competing risks environment. It is a generalization of Turnbull's (1976) classic univariate algorithm. The algorithm was published in Adamic et al. (2010) and Adamic & Caron (2014).

Usage

survCompeting(data, tau, n, nc, epsilon)

Arguments

data	input matrix of probabilities
tau	the vector of time points corresponding to columns in input matrix
n	the number of intervals in the dataset corresponding to rows in input matrix
nc	the number of causes (competing risks)
epsilon	small predermined value > 0

Value

Yj	estimated number at risk at time tau_j
djc	estimated number of events occuring at time tau_j by cause c
pjc	estimated probability for risk at time tau_j by cause c
djList	the list of d_j for every cause c
pjList	the list of p_j for every cause c
pjListold	the list of p_j for every cause c in the (iter - 1) iteration
iter	the number of iterations in the algorithm

Author(s)

Peter Adamic, Alicja Wolny-Dominiak

References

1. Adamic, P., Caron, S. (2014), "SC-CR Algorithms with Informative Masking", Scandinavian Actuarial Journal, 2014(4), 339-351.

2. Adamic, P., Dixon, S., Gillis, D. (2010), "Multiple Decrement Modeling in the Presence of Interval Censoring and Masking", Scandinavian Actuarial Journal, 2010(4), 312-327.

3. Adamic, P., Ouadah, S. (2009), "A Kernel Method for Modeling Interval Censored Competing Risks", South African Statistical Journal, 43(1), 1-20.

4. Turnbull, B. (1976). The Empirical Distribution Function with Arbitrarily Grouped, Censored and Truncated Data, Journal of the Royal Statistical Society. Series B (Methodological), 38(3), 290-295.

Examples

data(censoredMatrix)
df <- inputM(censoredMatrix)

res <- survCompeting(df$input, df$tau, 8, 3, 0.01)
res

#summary
round(res$Yj, 2)
round(res$djc, 2)
round(res$pjc, 2)
res$iter
sum(unlist(res$pjList))
sum(unlist(res$pjListold))

sccr documentation built on Dec. 15, 2020, 5:11 p.m.