TY - JOUR
T1 - Automated requirements engineering framework for agile model-driven development
AU - Umar, Muhammad Aminu
AU - Lano, Kevin
AU - Abubakar, Abdullahi Kutiriko
N1 - Publisher Copyright:
Copyright © 2025 Umar, Lano and Abubakar.
PY - 2025
Y1 - 2025
N2 - Introduction: Advances in requirements engineering, driven by various paradigms and methodologies, have significantly influenced software development practices. The integration of agile methodologies and model-driven development (MDE) has become increasingly critical in modern software engineering. MDE emphasizes the use of models throughout the development process, necessitating structured approaches for handling requirements written in natural language. Methods: This paper proposes an automated requirements engineering framework for agile model-driven development to enhance the formalization and analysis of textual requirements. The framework employs machine learning models to extract essential components from requirements specifications, focusing specifically on class diagrams. A comprehensive dataset of requirements specification problems was developed to train and validate the framework's effectiveness. Results: The framework was evaluated using comparative evaluation and two real-world experimental studies in the medical and information systems domains. The results demonstrated its applicability in diverse and complex software development environments, highlighting its ability to enhance requirements formalization. Discussion: The findings contribute to the advancement of automated requirements engineering and agile model-driven development, reinforcing the role of machine learning in improving software requirements analysis. The framework's success underscores its potential for widespread adoption in software development practices.
AB - Introduction: Advances in requirements engineering, driven by various paradigms and methodologies, have significantly influenced software development practices. The integration of agile methodologies and model-driven development (MDE) has become increasingly critical in modern software engineering. MDE emphasizes the use of models throughout the development process, necessitating structured approaches for handling requirements written in natural language. Methods: This paper proposes an automated requirements engineering framework for agile model-driven development to enhance the formalization and analysis of textual requirements. The framework employs machine learning models to extract essential components from requirements specifications, focusing specifically on class diagrams. A comprehensive dataset of requirements specification problems was developed to train and validate the framework's effectiveness. Results: The framework was evaluated using comparative evaluation and two real-world experimental studies in the medical and information systems domains. The results demonstrated its applicability in diverse and complex software development environments, highlighting its ability to enhance requirements formalization. Discussion: The findings contribute to the advancement of automated requirements engineering and agile model-driven development, reinforcing the role of machine learning in improving software requirements analysis. The framework's success underscores its potential for widespread adoption in software development practices.
KW - agile development
KW - machine learning
KW - model-driven development
KW - model-driven engineering
KW - NLP
KW - requirements engineering
UR - http://www.scopus.com/inward/record.url?scp=86000577885&partnerID=8YFLogxK
U2 - 10.3389/fcomp.2025.1537100
DO - 10.3389/fcomp.2025.1537100
M3 - Article
AN - SCOPUS:86000577885
SN - 2624-9898
VL - 7
JO - Frontiers in Computer Science
JF - Frontiers in Computer Science
M1 - 1537100
ER -