Package {pcsclr}

Title: Progressive Censoring Schemes with Competitive Latent-Risk
Version: 0.1.1
Maintainer: Okechukwu J. Obulezi <oj.obulezi@unizik.edu.ng>
Description: Implements simulation, numerical maximum likelihood estimation via fourth-order Runge-Kutta path optimization, and high-speed Bayesian Markov Chain Monte Carlo (MCMC) samplers for Weibull lifetimes under progressive censoring setups with competitive latent risks. Both point estimation and interval estimation are provided for the model parameters.
License: MIT + file LICENSE
Encoding: UTF-8
Imports: graphics, Rcpp (≥ 1.0.0), stats
LinkingTo: Rcpp, RcppArmadillo
Config/roxygen2/version: 8.0.0
NeedsCompilation: yes
Packaged: 2026-07-21 19:34:57 UTC; Dr. O. J. Obulezi
Author: Okechukwu J. Obulezi [aut, cre]
Repository: CRAN
Date/Publication: 2026-07-30 17:10:31 UTC

Fit PCS-CLR Parameter Estimation Models

Description

Fit PCS-CLR Parameter Estimation Models

Usage

fit_pcsclr(
  time,
  delta,
  removal,
  method = c("RK4", "Bayes_Kernel"),
  control = list(),
  is_hybrid = FALSE,
  t_max = 3,
  r_star = 0,
  ...
)

Arguments

time

Vector of observed lifetime intervals.

delta

Attrition risk indicator matrix.

removal

Active dynamic progressive censorship counts.

method

Selection between "RK4" paths or "Bayes_Kernel" sampler loops.

control

Optional configuration list parameters.

is_hybrid

Boolean indicator for hybrid censoring regimes.

t_max

Truncation time milestone boundary constraint.

r_star

Escaped survival components unfailed beyond cutoff point.

...

Additional structural arguments passed to internal optimization engines.

Value

An object of S3 class "pcsclr_fit" consisting of a list with the following elements:

estimates

A numeric vector containing the estimated Weibull shape (\alpha) and scale (\beta) parameters.

method

A character string indicating the algorithm used ("RK4" or "Bayes_Kernel").

iterations

Number of iterations to reach convergence (returned only when method = "RK4").

chains

A list of posterior MCMC samples for alpha and beta parameters (returned only when method = "Bayes_Kernel").

Examples

# 1. Generate small sample mock variables
set.seed(123)
n <- 30
mock_time <- rweibull(n, shape = 1.5, scale = 2.0)
mock_delta <- sample(c(0, 1), n, replace = TRUE, prob = c(0.2, 0.8))
mock_removal <- rbinom(n, 5, 0.1)

# 2. Fit using 4th-order Runge-Kutta Optimization
res_rk4 <- fit_pcsclr(time = mock_time, delta = mock_delta,
                      removal = mock_removal, method = "RK4")
print(res_rk4$estimates)

# 3. Fit using the C++ Bayesian MCMC Sampler
res_bayes <- fit_pcsclr(time = mock_time, delta = mock_delta,
                         removal = mock_removal, method = "Bayes_Kernel",
                         control = list(M = 1000, burn_in = 200))
print(res_bayes$estimates)

Simulate Datasets Under PCS-CLR Configurations

Description

Simulate Datasets Under PCS-CLR Configurations

Usage

sim_pcsclr(
  n,
  target,
  alpha = 1.8,
  beta = 2.5,
  p = 0.15,
  scheme = c("baseline", "hybrid"),
  cs_layout = c("CS_I", "CS_II"),
  T_max = 3
)

Arguments

n

Total initial sample population count.

target

Scheduled tracking truncation parameter target threshold.

alpha

Shape vector coefficient.

beta

Scale distribution tracker.

p

Attrition or competitive latent risk probability.

scheme

Setting options: "baseline" or "hybrid".

cs_layout

Distribution setups: "CS_I" or "CS_II".

T_max

Explicit global temporal limitation window boundary.

Value

An object of S3 class "pcsclr_data" consisting of a list with the following elements:

T_obs

A numeric vector of observed failure times up to the target or truncation cutoff.

delta

A numeric vector of risk indicators corresponding to each observed time.

R_vector

A numeric vector of progressive censoring removal counts at each failure point.

scheme

A character string indicating the active censoring scheme ("baseline" or "hybrid").

is_truncated

A logical flag indicating whether time truncation (T_max) occurred during simulation.

T_max

The temporal limit applied (returned only when is_truncated = TRUE).

R_star

The count of remaining survival components unfailed beyond the truncation cutoff (returned only when is_truncated = TRUE).