Verification method of requirements documents with a natural language

2023 Fiscal Year Final Research Report

• Project/Area Number: 19K11913
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 60050:Software-related
• Research Institution: Ritsumeikan University
• Principal Investigator: Ohnishi Atsushi 立命館大学, 情報理工学部, 教授 (50160560)
• Project Period (FY): 2019-04-01 – 2024-03-31
• Keywords: 日本語要求の抽出と検証

Outline of Final Research Achievements

In this research, we focus on functional requirements, usability requirements, response requirements, reliability requirements, maintenance requirements and security requirements. We propose a derivation method of these requirements from a software requirements document and a verification method of the completeness, the consistency, the non-redundancy and the unambiguity of derived requirements.
We established a derivation method of these requirements using keywords and a verification method of derived requirements, developed a prototype system based on the methods, and evaluated both the methods and the system. We also proposed a derivation method of the requirements with generative AI in order to increase the quality of the method.

Free Research Field

ソフトウェア工学

Academic Significance and Societal Importance of the Research Achievements

失敗に終わった456件のソフトウェア開発プロジェクトにおいて、半数近くに当たる51.5%のプロジェクトで要求（要件）定義の失敗を原因とした報告があり、要求定義やその成果物である要求仕様書の正確さが求められている。
本研究で提案する要求仕様書や要求文書の誤り検出手法を適用することにより、システム開発プロジェクトの生産性や品質を向上させると共に、プロジェクトの失敗を防ぐことが可能となり、その社会的な意義は少なくないと考えられる。