Software Reliability Assessment Based on a Discretized Model by Bayes' Theory

摘要

We discuss an interval estimation approach for model parameters and software reliability assessment measures of a discretized software reliability model, which has a consistency with an actual collecting activity of discrete software fault-count data and is expected to conduct highly-accurate assessment of software reliability. The interval estimation must be useful for conducting software reliability assessment because most of software reliability data are incomplete. Concretely speaking, we derive conditional probability distributions of model parameters by following the Bayes' theory. Then, we apply the Gibbs sampling method, which is one of the Markov chain Monte Carlo (MCMC) method, for sampling the model parameters from the derived conditional probability distributions of model parameters. The probability distributions of model parameters can be obtained through this procedures. Further, this paper shows numerical examples of our approach by using actual fault count data.