An automated framework for prioritizing software requirements

Requirement Engineering (RE) is a critical phase in software development, integral to the successful execution of projects. The initial stage of RE involves requirement elicitation and analysis, where the prioritization of requirements is critical. Traditional methods of requirement prioritization (RP) are diverse, each presenting unique challenges. In response to the challenges of traditional methods, this paper proposes an entirely automated framework designed to eliminate the disadvantages associated with excessive stakeholder involvement. This innovative framework processes raw natural language inputs directly, applying a three-phase approach to systematically assign priority numbers to each requirement. The first phase preprocesses the input to standardize and prepare the data, the second phase employs advanced machine learning algorithms to analyze and rank the requirements, and the third phase consolidates the results to produce a final prioritized list. The effectiveness of this method was tested using the RALIC (Replacement Access, Library, and ID Card) dataset, a well-known benchmark in the field of requirement engineering. The results confirm that our automated approach not only enhances the efficiency and objectivity of the prioritization process but also scales effectively across diverse and extensive sets of requirements. This framework represents a significant advancement in the field of software development, offering a robust alternative to traditional, subjective methods of requirement prioritization.