README.md

In nathansam/SMLN: Bayesian Survival Analysis Using Shape Mixtures of Log-Normal Distributions

BASSLINE

| Usage | Release | Development | |-------------------------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | | | | | | | | | | | |

Overview

Mixtures of life distributions provide a convenient framework for survival analysis: particularly when standard models such as the Weibull or the log-normal are unable to capture some features from the data. These mixtures can also account for unobserved heterogeneity or outlying observations.

BASSLINE (BAyeSian Survival anaLysIs usiNg shapE mixtures of log-normal distributions) uses shape mixtures of log-normal distributions to fit data with fat tails and has been adapted from code written by Vallejos & Steel[1]. Some of the functions have been rewritten in C++ for increased performance.

5 distributions from the log-normal family are supported by BASSLINE:

• The log-normal distribution
• The log student's T distribution
• The log-logistic distribution
• The log-Laplace distribution
• The log-exponential power distribution

As well as MCMC (Markov chain Monte Carlo) algorithms for the 5 distributions, additional functions which allow log-marginal likelihood estimators and deviance information criteria to be calculated are provided. Case deletion analysis and outlier detection are also supported.

Installation

BASSLINE is currently not available on CRAN but can be installed via the devtools package

devtools::install_github("nathansam/BASSLINE")

References

• [1] Vallejos & Steel (2015). Journal of the American Statistical Association.

nathansam/SMLN documentation built on May 14, 2022, 9:07 p.m.