Abstract: Interval estimation in assessment in software systems must be useful because it is hard to obtain a sufficient amount of software reliability data for conducting point estimation and the software reliability data is essentially incomplete. We discuss flexible software reliability measurement considering the uncertainty of model parameters in a reliability model. Concretely, applying a discrete-time domain model, we analyze the Bayesian inference of the parameters in the model by using the Markov chain Monte Carlo method. Furthermore, numerical illustrations of the approach for our method are also shown in this paper.

Key words: software reliability assessment, software reliability model, integrable difference equation, Bayesian interval estimation, MCMC method