Log-Concave Distribution Estimation with Interval-Censored Data.

iclogcondist: Nonparametric Estimation for a Log-concave Distribution Function with Interval-censored Data in R

Overview

The iclogcondist package provides an efficient algorithm to compute the nonparametric maximum likelihood estimator (NPMLE) of a log-concave distribution function for the underlying event time in mixed-case interval-censored data. The computational algorithm combines an active set method with an iterative convex minorant algorithm.

Features

• Distribution Function Estimation: The package provides a function ic_LCMLE to estimate the distribution function assuming its log-concavity, as well as functions to compute the unconstrained MLE and the distribution function corresponding to the least concave majorant (LCM) of the logarithm of the unconstrained MLE.

• Simulation Capabilities: Users can simulate interval-censored datasets where the underlying event times follows some common parametric families with a log-concave distribution function.

• Visualization: The package includes plotting functions to compare the estimated distribution functions obtained from the log-concave MLE, the unconstrained MLE, and the distribution function corresponding to the LCM of the logarithm of the unconstrained MLE.

Installation

You can clone the repository iclogcondist locally, by running this command in your terminal

git clone https://github.com/ChaoyuYuan/iclogcondist.git

Option 1: Build and Install Locally

If you have cloned the repository locally, follow these steps to build and install the package manually:

1. Build the package:

In R, navigate to the package directory and build the package:

setwd("path/to/iclogcondist")  # Change to your package directory
devtools::document()     # Generate documentation
devtools::build()        # Build the package

2. Install the package:

After building, you will get the package object iclogcondist_1.0.0.tar.gz. You can install the package locally:

install.packages("path/to/iclogcondist_1.0.0.tar.gz", repos = NULL, type = "source")

or

devtools::install_local("path/to/iclogcondist_1.0.0.tar.gz")

The second command will install the prerequisite packages, such as Rcpp and ggplots, automatically during the process, while the first command will not.

Option 2: Install Directly from Local Files

You can also install the package directly from the local directory without building manually:

devtools::install("path/to/iclogcondist")

Similarly, this command will install the prerequisite packages automatically.

Loading the Package

Once the package is installed, load it with:

library(iclogcondist)

Usage Example

Here is a simple example of how to use the iclogcondist package:

data(lgnm)
X <- lgnm

# Run the LC MLE algorithm
result <- ic_LCMLE(X)

# Print the estimated distribution function
print(result$est$F_hat)

For more details in usage examples, please refer to the iclogcondist_Example.pdf file.

License

This package is licensed under the GPL-3 License.

Contributing

Contributions are welcome! If you would like to report issues or contribute to the development, please feel free to open an issue or submit a pull request.