Home // International Journal On Advances in Software, volume 18, numbers 3 and 4, 2025 // View article

Design Pattern Detection in Source Code: A Neural Graph Database Approach as an Ensemble Base Model

Authors:
Roy Oberhauser
Peter Schneider

Keywords: software design pattern detection; machine learning; neural graph databases; graph neural networks; ensemble methods; software design patterns; software engineering

Abstract:
Software design patterns offer reusable structural solutions that support developers and maintainers in addressing common design problems. Their abstractions can support program code documentation and comprehension, yet manual pattern documentation via code or code-related artifacts (documents, models) can be unreliable, incomplete, and labor-intensive. Various automated Design Pattern Detection (DPD) techniques have been proposed, yet adoption remains limited and further investigation of viable solutions is needed. Towards more effective automated DPD, this paper contributes our Neural Graph Database approach DPD-NGDB, which also functions as a base model in our Ensembles Methods approach DPD-EM. The realization demonstrates the feasibility of our approaches, while the evaluation compares and benchmarks the DPD performance against a Gang-of-Four (GoF) software design pattern dataset, demonstrating its potential.

Pages: 106 to 126

Copyright: Copyright (c) to authors, 2025. Used with permission.

Publication date: December 30, 2025

Published in: journal

ISSN: 1942-2628