摘要：The quality of software products rarely achieves the level expected by managers purchasing the software. Only one of five implemented software projects successfully fulfills the required quality criteria. As with other products, customer satisfaction is the basic quality criterion for software products. Because each software product must be designed for an individual user, neither the customer nor the software producer has an exact image of the final product or its quality. Even when the customer and producer have similar images, the product is subject to change during the software development lifecycle. Customers often modify software requirements as the underlying needs of the customer change. It is not possible to generalize the relative importance of software quality characteristics because they are unequal for different products as well as for different customers. The aim of this research is to explore the possibilities of knowledge-based systems to support and automate the process of identifying important software quality characteristics at a specific moment in time. Identifying quality characteristics will facilitate software quality management. We applied a modified C4.5 algorithm and induced the most informative attributes with rules describing a variety of perceptions of software quality characteristics from customer and producer perspectives. The findings provide managers with useful insights that will signal needs for software quality improvements, which will enhance and refine their business performance.
