A theory of software reliability and its application
IEEE Transactions on Software Engineering
Hi-index | 0.00 |
In the development of large-scale computer software and in the management of the development process, it is often useful to model the reliability and the cost of development of these software packages. There have been many papers that develop models and show their usefulness as management tools. The models that use Bayesian methodology assume that a prior distribution is given. Our paper offers a methodology of assessing a prior distribution subjectively. Two computer programs have been developed for this particular purpose: One assesses a subjective prior distribution and the other suggests a family of probability functions. The importance of consistent prior distributions is twofold. First, these distributions reflect consistent initial predictions because they are developed by a structured process. Second, these distributions are the starting point for applying Bayes' theorem to develop the posterior distribution by modifying the prior distribution with actual data available later.