An integrated neural-fuzzy system of software reliability prediction

Abstract

The paper explores the potential of prediction techniques which have been used for assessing software reliability. It is demonstrated that artificial neural networks (ANN) provide the basis for fully satisfactory prediction methods, which are shown to be superior to conventional analytical models in many respects. The method retains all the advantages normally associated with its use in other applications, and brings significant merits in this prediction study. The basic mode of computer algorithm is presented. The incorporation of fuzzy logic systems in two strategic blocks leads to the achievement of more efficient and decisive system of software reliability prediction.