Application of Machine Learning Paradigms for Predicting Quality in Upstream Software Development Life Cycle

Prediction and estimation of the software quality early in the life cycle of software development have become an increasingly important problem. The primary factors affecting the determination of the software quality are the nature of the development process and the specification of the product in the upstream development phases. Current trends in software engineering are promoting the idea of processes maturity which improves the software quality. Models for integrating the product and process attributes are the need of the hour to ensure that process improvement actions are going into the right direction and software quality is improved. As a step towards building such model, this paper examines the application of machine learning paradigms like Artificial Neural Network, Case Based Reasoning, Rule Induction and Genetic Algorithm towards predicting the software quality characteristics and identifies the potential applications for further research.