A Design and Development of Online Quality Evaluation System for Natural Language Based Software Requirements

• Main Authors: Chia-Hui Liu, 劉佳惠
• Other Authors: Chung-Yang Chen
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/38qd7j

碩士 === 國立中央大學 === 資訊管理學系 === 105 === Nowadays, as software project development life cycle becomes shorter and shorter, retaining quality of software project development has become a critical issue among organizations. By maintaining quality of software project development at early phase of requirement elicitation, organizations can both decrease development-related cost and increase efficiency in software project development. Furthermore, plenty of experts have proposed different methods to evaluate the quality of natural language based requirement to have more complete UML diagrams at early phase of software development; that is, system analysis. On the other hand, if members in an organization can share their knowledge of requirement evaluation with other members so that they can utilize it, organization will have better efficiency in requirement evaluation in the future. Hence, in this study, we propose an online evaluation system – Online Requirement Evaluation and Knowledge Sharing System (OREKSS), which aims at evaluating completeness of nature language requirement from relative stakeholders. In addition, completeness is consisted of three quality dimensions – unambiguity, necessity, and reachability. After relative stakeholders input their requirement description, OREKSS will provide index of unambiguity, necessity, reachability, and completeness as well as recommendation to make requirement more complete. Moreover, knowledge engineers can improve evaluation rules of OREKSS so that requirement evaluation results can meet relative stakeholders’ needs. Therefore, in software project development, OREKSS not only augments efficiency but also reduce cost at early phase by evaluating requirement completeness.