Description

Title Predicting software reliability using Bayesian belief networks.
Abstract Predicting software reliability during early design specification helps to prevent failures and costly refactoring later on. I propose to semi-automatically map failure dependencies between interactions in UML sequence diagrams to a reliability prediction model in a Bayesian belief network. The approach was evaluated in two industrial case studies. During this WAI presentation I want to tell you about above research and welcome your feedback on the approach and findings.