An Automated Requirement Ranking Approach for Identifying Software Requirement

Objective: This paper presents an automated requirement ranking approach for identifying software requirements, which combines the project requirements of the order of approximation calculated through learning techniques. Method: We proposed an algorithm RRA, which automatically calculates, approximate ranks for the requirements based on priority rules. Findings: The algorithm considers requirements as inputs and outputs the best suitable requirements for software development in three stages namely Pairing of requirements, Extraction of priority and Learning of Priority. The proposed algorithm is more effective compared to CBRank, especially for more number of extracted pairs. Improvement: The work can be extended to produce accurate ranks for more number of extracted requirement pairs.