3. DESCRIPTION OF PROGRAM OR FUNCTION
BURD (Bayesian Update for Reliability Data) is a simple code that can be used to obtain a Bayesian estimate easily in the data analysis of PSA (Probabilistic Safety Assessment). According to the Bayes' theorem, basically, the code facilitates calculations of posterior distribution given the prior and the likelihood (evidence) distributions.
The distinctive features of the program, BURD, are the following:
- The input consists of the prior and likelihood functions that can be chosen from the built-in statistical distributions.
- The available prior distributions are uniform, Jeffrey's non informative, beta, gamma, and log-normal that are most-frequently used in performing PSA.
- For likelihood function, the user can choose from four statistical distributions, e.g., beta, gamma, binomial and poisson.
- A simultaneous graphic display of the prior and posterior distributions facilitate an intuitive interpretation of the results.
- Export facilities for the graphic display screen and text-type outputs are available.
- Three options for treating zero-evidence data are provided.
- Automatic setup of an integral calculus section for a Bayesian updating.