Requirement prioritisation and selection is an essential activity in modern-day large software development. Optimal prioritisation process is critical for successful implementation and release planning in a software development project. Requirement prioritisation becomes more challenging in projects having large sets of requirements and stakeholders, having diverse perspectives resulting in irrelevancy and ambiguity during features extraction. This study aims to improve requirement prioritisation process using text mining and clustering techniques for accurate extraction of features and requirement prioritisation in multi-stakeholder context. The proposed framework developed to avoid incompleteness in requirements and disagreement among development teams and stakeholders. Thus, the proposed framework compared with other requirement prioritisation techniques (i.e. Analytical Heretical Process, Commutative Voting and Wiegers) to highlight the significance of the proposed framework while conducting an experimental study. The results show that the proposed framework outperformed the traditional techniques and enhanced the prioritisation process with complete semantic information of extracted features and taking into account the diverse perspective of stakeholders.

中文翻译：

---

基于方面的需求挖掘技术，用于改进优先级过程：多利益相关方的观点

在现代大型软件开发中，需求优先级和选择是一项必不可少的活动。最佳优先级排序过程对于软件开发项目中成功的实施和发布计划至关重要。在具有大量需求和利益相关者的项目中，需求优先级变得更具挑战性，因为项目的观点多种多样，导致特征提取期间不相关且含糊不清。这项研究旨在使用文本挖掘和聚类技术来改进需求优先级排序过程，以在多利益相关方环境中准确提取特征和需求优先级。拟议框架的制定是为了避免开发团队和利益相关者之间需求的不完整和分歧。从而，提出的框架与其他需求优先级排序技术（即，分析异端过程，可交换投票和Wiegers）进行了比较，以在进行实验研究时突出提出的框架的重要性。结果表明，该框架优于传统技术，并利用提取特征的完整语义信息并考虑了利益相关者的不同观点，从而增强了优先级排序过程。