| Title: | Progressive Regularized Vine Copula for Masked Competing Risks |
|---|---|
| Description: | Implements Progressive Regularized Vine Copula (Prog-Vine) frameworks for high-dimensional dependent competing risks with masked failure causes under Progressive Type-II Censoring. Fits Weibull marginals, estimates pair-copula trees using Expectation-Maximization (EM) algorithms, computes Louis observed information confidence intervals, and implements Data Augmentation Gibbs Samplers for Bayesian credible intervals. |
| Authors: | Okechukwu J. Obulezi [aut, cre] |
| Maintainer: | Okechukwu J. Obulezi <[email protected]> |
| License: | GPL-3 |
| Version: | 0.1.0 |
| Built: | 2026-10-09 11:28:41 UTC |
| Source: | https://github.com/cran/ProgVine |
Weibull Density for Progressive Censoring
dweibull_prog(t, alpha, beta)dweibull_prog(t, alpha, beta)
t |
Numeric vector of failure times. |
alpha |
Numeric scale parameter (> 0). |
beta |
Numeric shape parameter (> 0). |
Numeric vector of density values.
dweibull_prog(c(1.2, 2.5), alpha = 2.0, beta = 1.5)dweibull_prog(c(1.2, 2.5), alpha = 2.0, beta = 1.5)
Implements Algorithm 5 (EM-Vine MLE Engine) for high-dimensional dependent competing risks with masked causes under Progressive Type-II Censoring.
fit_em_vine(t, C, R, d = 2, max_iter = 50, tol = 1e-05)fit_em_vine(t, C, R, d = 2, max_iter = 50, tol = 1e-05)
t |
Numeric vector of failure times. |
C |
Matrix or list of masking sets (binary matrix or indicator list). |
R |
Numeric vector of progressive removal counts. |
d |
Integer specifying number of competing risks. |
max_iter |
Integer specifying maximum EM iterations. |
tol |
Numeric tolerance convergence criterion. |
List containing estimated Weibull parameters, weights, and log-likelihood.
set.seed(42) n <- 20 t_sim <- runif(n, 0.5, 5.0) R_sim <- rep(0, n) C_sim <- matrix(sample(c(0, 1), n * 2, replace = TRUE, prob = c(0.3, 0.7)), ncol = 2) fit <- fit_em_vine(t = t_sim, C = C_sim, R = R_sim, d = 2, max_iter = 10) print(fit$alpha)set.seed(42) n <- 20 t_sim <- runif(n, 0.5, 5.0) R_sim <- rep(0, n) C_sim <- matrix(sample(c(0, 1), n * 2, replace = TRUE, prob = c(0.3, 0.7)), ncol = 2) fit <- fit_em_vine(t = t_sim, C = C_sim, R = R_sim, d = 2, max_iter = 10) print(fit$alpha)
Implements Algorithm 6 for Prog-Vine Bayesian Inference.
gibbs_prog_vine(t, C, R, d = 2, n_mc = 100, burn_in = 20)gibbs_prog_vine(t, C, R, d = 2, n_mc = 100, burn_in = 20)
t |
Numeric vector of failure times. |
C |
Binary matrix indicating cause masking sets. |
R |
Vector of progressive removals. |
d |
Integer number of competing risks. |
n_mc |
Total MCMC samples. |
burn_in |
MCMC burn-in length. |
List of posterior chains for alpha and beta parameters.
set.seed(42) t_sim <- runif(15, 0.5, 3.0) R_sim <- rep(0, 15) C_sim <- matrix(1, nrow = 15, ncol = 2) gibbs_res <- gibbs_prog_vine(t_sim, C_sim, R_sim, d = 2, n_mc = 30, burn_in = 5) summary(gibbs_res$alpha_chain)set.seed(42) t_sim <- runif(15, 0.5, 3.0) R_sim <- rep(0, 15) C_sim <- matrix(1, nrow = 15, ncol = 2) gibbs_res <- gibbs_prog_vine(t_sim, C_sim, R_sim, d = 2, n_mc = 30, burn_in = 5) summary(gibbs_res$alpha_chain)
Weibull Cumulative Distribution for Progressive Censoring
pweibull_prog(t, alpha, beta)pweibull_prog(t, alpha, beta)
t |
Numeric vector of failure times. |
alpha |
Numeric scale parameter (> 0). |
beta |
Numeric shape parameter (> 0). |
Numeric vector of cumulative probability values.
pweibull_prog(c(1.2, 2.5), alpha = 2.0, beta = 1.5)pweibull_prog(c(1.2, 2.5), alpha = 2.0, beta = 1.5)