Development and Evaluation of Automatic Summarization Tool for Requirements Specification

2023 Fiscal Year Final Research Report

• Project/Area Number: 19K11907
• 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: Kogakuin University
• Principal Investigator: Mari Inoki 工学院大学, 情報学部(情報工学部), 教授 (10739634)
• Project Period (FY): 2019-04-01 – 2024-03-31
• Keywords: 要求工学 / 要求仕様書 / 自動要約 / 理解支援 / 記述状況可視化 / 自然言語処理 / AI

Outline of Final Research Achievements

The requirements specification document includes the background, purpose, functional requirements, and non-functional requirements of target systems. We propose an automatic summarization support tool for these specifications. For functional requirements, our tool extracts design elements, while for non-functional requirements, it provides a risk summary indicating potential definition omissions. The background and purpose are summarized using a conventional method to extract key sentences. Additionally, we offer a method to support document comprehension by indicating where important concepts are located within the requirements specification. Evaluation confirmed that this method effectively accelerates engineers' understanding of the requirements specification. Furthermore, visualizing the description status in various technical documents is expected to be useful for knowledge transfer to engineers.

Free Research Field

ソフトウェア工学

Academic Significance and Societal Importance of the Research Achievements

Digital Transformationへの取り組みが活発化し、技術者にとり、情報システムの要求仕様書の作成に加え、関連するシステムの要求仕様書等の技術文書を参照理解する機会が増加している。開発した自動要約ツールを活用して、要求仕様書の要約情報を用いることで、新規ビジネスアイデア創出、リスク予測、基本設計の効率化、仕様変更の影響箇所の特定の支援が可能になる。本技術を大規模かつ高度な内容の技術文書に活用し、要約情報を用いて、知識を形式知化・共有することで、組織の知識継承にも貢献し、IT人材不足の解決やDigital Transformationの社会実装の一層の加速への貢献が期待できる。