A Recommender System based on Intuitionistic Fuzzy Sets for Software Requirements Prioritization

Abstract

Requirements Prioritization (RP) is an important activity in requirements engineering aiming to give priority and order to requirements for implementation in the next version of a software project. RP is applied iteratively, according to various prioritization criteria, by multiple project stakeholders who may have different roles, needs and knowledge. In large-scale software projects, where the set of candidate requirements is large, stakeholders may not be interested in evaluating all requirements and they may not have the expertise, the time or the willingness to consider all candidate requirements. Recommender Systems (RS) can be a useful solution to information overload when stakeholders have to evaluate a large number of alternatives. During evaluation of requirements, it is often practically impossible to ensure that all stakeholders have complete knowledge on all requirements. Thus, stakeholders may show some degree of uncertainty and hesitation, as it is difficult to precisely evaluate each requirement according to each prioritization criterion. The intuitionistic Fuzzy Sets (IFSs) are an extension of fuzzy sets which can deal with stakeholders' uncertainty and hesitation regarding the prioritization criteria importance and requirements ratings. The aim of this paper is to present an RS approach based on the collaborative filtering technique to effectively provide suggestions to stakeholders while prioritizing requirements. The proposed RS approach is tested using a publicly available large dataset of software requirements and the results show an improved performance. © 2021 ACM.