Bayesint: A Python Package for Calculating Bayesian Credible Intervals of Ratios of Beta Distributions. (21st December 2021)

• Record Type: Journal Article
• Authors: Bekker-Nielsen Dunbar, M.; Finnie, Thomas J. R.
• Is Part Of: Journal of open research software. Volume 9(2021)
• Publication Date: 2021-12-21
• Subjects: beta distributions -- cholera -- contingency tables -- credible interval -- relative risk -- Python; Computer software -- Reusability -- Periodicals; Open source software -- Periodicals; 005
• DOI: 10.5334/jors.283
• Languages: English
• ISSNs: 2049-9647

Abstract

We present bayesint, a Python package for calculating Bayesian credible intervals for ratios of two independent beta distributions such as are used when considering binomial data. Such data could be found in counts of events in medical diagnostic tests, case-control studies, and field epidemiological investigations, among other applications. The package contains general functions providing the expression of the density and distribution of the ratio and two functions for calculating the equal-tailed and highest posterior density credible intervals. The package is intended for use with 2×2 contingency tables. We introduce bayesint by comparing two groups in a contingency table through calculating the relative risk of cholera, showcasing its use in a novel context.