Generating Datasets for Software Requirements Prioritization Research

Abstract

Software requirements are the need of the stakeholders which are identified during requirements elicitation process. There are different types of the software requirements like functional requirements, non-functional requirements, testing requirements, etc. After requirements elicitation process, a system may have hundreds or thousands of requirements; and during the development process it is not possible to implement all the software requirements because of the several limitations of an organization as well as the stakeholders. To develop the successful software product, the elicited software requirements must be prioritized before the development of software so that high ranked software requirements can be identified. Computation of the ranking values of the software requirements is a difficult task because it involves lot of computational steps like evaluation of software requirements, computation of group preferences from the individual preferences, computation of the ranking values after normalizing the preference matrices, etc. Among these steps, in this paper, we mainly focus on the evaluation of the software requirements using linguistic variables because in real life applications people use linguistic variables to specify their preferences on software requirements. The objective of this paper is to generate a dataset of 32 functional requirements after evaluating it on the basis of 3 non-functional requirements so that the generated dataset can be used as an input into the software requirements prioritization methods. The institute examination system is used as case study to explain the steps for the generation of the dataset.