earticle

영어

Agile development is truly the need of the hour due to its numerous advantages which are in line with the present business trends. A successful requirement engineering serves as a foundation for success for any software development project. Functional requirements point towards the product services and non-functional requirements are related to the emergent properties of the system. Correct and speedy elicitation of functional and non-functional requirements contribute a great deal towards successful requirement engineering process. Many techniques have been proposed in the past for requirement elicitation for agile development, but they do not take into consideration a holistic automatic approach concerning functional and non-functional requirements. This paper proposes a supervised learning based automated (neural network with the genetic algorithm) approach for successfully classifying functional and non-functional requirements from multiple requirements documents in an agile environment. It is implemented on two data sets and further analysis, and comparison of this model is made with an another implemented model (SVM with RBF kernel) based on precision, recall, and accuracy. This paper contributes in simplifying and automating the requirement engineering process; thus, making the life easier for many stakeholders.