A Neuro-Fuzzy based Approach to Software Quality Requirements Prioritization

Requirements engineering is one of the important topics in software engineering and correct understanding of requirements can affect the total quality of the whole software system. Lack of considering or wrong identification of requirements not only leads to customers' dissatisfaction, but also leads to the software failure and increases the costs of development. Requirement prioritizing plays an important role in decision making process in subsequent phases and eliminates the complexities resulting from vague requirements. The aim is to find the right order of main requirements. Neuro-fuzzy system combines the ability of human reasoning (logic) through predetermined laws with the learning ability of neural networks and can apply the human expertise to find the total output value by using a set of laws stored in knowledge base. In this paper, we propose an effective approach to prioritizing software quality requirements based on neuro-fuzzy system. This paper shows that our proposed approach is more efficient in terms of both the time consumption and the learning ability compared to related approach.