| Title: | Simulate and Analyze Interval- and Mixed-Censored Survival Data |
| Version: | 0.1.0 |
| Description: | Provides tools to simulate and analyze survival data with interval-, left-, right-, and uncensored observations under common parametric distributions, including "Weibull", "Exponential", "Log-Normal", "Log-Logistic", "Gamma", "Gompertz", "Normal", "Logistic", and "EMV". The package supports both direct maximum likelihood estimation and imputation-based methods, making it suitable for methodological research, simulation benchmarking, and teaching. A web-based companion app is also available for demonstration purposes. |
| License: | MIT + file LICENSE |
| Encoding: | UTF-8 |
| RoxygenNote: | 7.3.2 |
| URL: | https://github.com/jayarasan/simIC |
| BugReports: | https://github.com/jayarasan/simIC/issues |
| NeedsCompilation: | no |
| Packaged: | 2025-07-29 04:30:24 UTC; user |
| Author: | Jayanthi Arasan [aut, cre] |
| Maintainer: | Jayanthi Arasan <jayanthi@upm.edu.my> |
| Repository: | CRAN |
| Date/Publication: | 2025-07-31 10:00:02 UTC |
Imputation-Based MLE for Censored Data
Description
Estimates distribution parameters using imputed event times.
Usage
mle_imp(
left,
right,
dist = "weibull",
impute = c("midpoint", "random", "median", "harmonic_median", "geometric_median",
"random_survival")
)
Arguments
left |
Left bounds of censoring intervals |
right |
Right bounds of censoring intervals |
dist |
Distribution name (e.g. "weibull", "loglogistic", "EMV") |
impute |
Imputation method: "midpoint", "random", "median", "harmonic_median", "geometric_median", "random_survival" |
Value
A list containing estimates, standard errors, and log-likelihood
Examples
# Simulate interval-censored data from a Weibull distribution
set.seed(123)
dat <- simIC(n = 100, dist = "weibull", shape = 1.5, scale = 5, width = 2,
study_start = 1, study_end = 8, uncensored_tol = 0.1)
# Fit model using harmonic median imputation
fit <- mle_imp(left = dat$left, right = dat$right, dist = "weibull", impute = "harmonic_median")
# Inspect results
print(fit$estimates)
print(fit$logLik)
print(fit$converged)
Interval-Censored Maximum Likelihood Estimation
Description
Estimates distribution parameters by maximizing the interval-censored likelihood.
Usage
mle_int(left, right, dist)
Arguments
left |
Left bounds of censoring intervals |
right |
Right bounds of censoring intervals |
dist |
Distribution name (e.g. "weibull", "loglogistic", "EMV") |
Value
A list containing estimates, standard errors, log-likelihood, and convergence status
Examples
# Simulate data from a log-logistic distribution
set.seed(123)
data <- simIC(n = 100, dist = "loglogistic", shape = 1.5, scale = 5, width = 2,
study_start = 1, study_end = 8, uncensored_tol = 0.1)
# Fit the model
fit <- mle_int(left = data$left, right = data$right, dist = "loglogistic")
print(fit$estimates)
print(fit$logLik)
print(fit$converged)
Simulate Interval-, Left-, Right-, and Uncensored Survival Data
Description
Simulates survival data with optional left-censoring, right-censoring, and uncensoring thresholds.
Usage
simIC(
n = 100,
dist = "weibull",
shape = 2,
scale = 1,
meanlog = 0,
sdlog = 1,
location = 0,
width = 1,
visit_start = 0,
study_start = NULL,
study_end = NULL,
uncensored_tol = 0.1
)
Arguments
n |
Number of samples. |
dist |
Distribution name ("weibull", "exp", "lognormal", "loglogistic", "normal", "logistic", "EMV", "gamma", "gompertz"). |
shape, scale |
Distribution parameters for applicable distributions. |
meanlog, sdlog |
For lognormal. |
location |
For normal, logistic, and EMV. |
width |
Visit interval width. |
visit_start |
First visit time. |
study_start |
Optional: left-censoring cutoff. |
study_end |
Optional: right-censoring cutoff. |
uncensored_tol |
Tolerance to treat (left, right) as exact event. |
Value
A data frame with columns: id, left, right, true_time, censoring
Examples
# Simulate 100 survival times from a log-normal distribution
set.seed(123)
data <- simIC(n = 100, dist = "lognormal", meanlog = 1.5, sdlog = 0.5, width = 2,
study_start = 1, study_end = 8, uncensored_tol = 0.1)
head(data)