Thesis (M.Sc.) - Cairo University - Faculty of Graduate Studies for Statistical Research - Department of Computer and Information Science

For the management of a software development project, an early software defect prediction is very helpful since it can provide the management with useful information about the target software quality. Many software defect prediction techniques and models have been proposed in the literature. But, the complicated situations of software development process call for a model that can consider all factors that have an impact on the quality of the target software. A software defect prediction model based on a fuzzy neural network that called ANFIS (Adaptive Neuro Fuzzy Inference System) have been proposed in the latest studies. This model is a hybrid model of Artificial Neural Network (ANN) and Fuzzy Logic (FL), which exploits the advantages of ANN and FL while eliminating their limitations, so it has the ability to handle the complicated non-linear relationships between these factors that have impact on software quality under a certain level of uncertainty. The main objective of this research is to enhance that ANFIS model, the proposed enhancements are complexity reduction and interpretability improvement of that neuro-fuzzy defect prediction model. The experimental results show that the proposed enhancements have significantly improved the model interpretability and reduced its complexity by reducing the number of the model rule-base by about 70% without affecting the model's performance except by a negligible value

Issued also as CD