PATCHES : A framework for contextualizing ADHD symptoms manifestations

Traore, Ibrahim and Tawil, Abdel-Rahman H. and Vlachos, Konstantinos and Belmabrouk, Yassine and Megdiche, Imen and Marquet-Doléac, Jérôme and Chaari, Lotfi (2026) PATCHES : A framework for contextualizing ADHD symptoms manifestations. Expert Systems With Applications, 331. p. 133339. ISSN 0957-4174

Text PATCHES.pdf - Accepted Version Restricted to Repository staff only until 18 June 2028. Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (5MB) | Request a copy

Abstract

Attention Deficit Hyperactivity Disorder (ADHD) presents substantial challenges for diagnosis and management due to its highly context-dependent manifestations across school, home, and social environments. Existing research on contextual modeling and ontology-based reasoning has offered valuable foundations for structuring knowledge, yet these efforts typically rely on static rule sets, overlook patient-specific variability, and lack scalable mechanisms for ontology population. Furthermore, many approaches remain at the level of conceptual design or limited experimental validation, without demonstrating practical clinical or educational relevance. To overcome these limitations, we introduce PATCHES, a modular ontology-based framework designed to dynamically contextualize ADHD symptom manifestations. PATCHES integrates patient health profiles with contextual factors such as activities, environment, time, and social presence enabling fine-grained situation interpretation. A key innovation is a multi-agent population pipeline, which substantially improves the precision, recall, and robustness of ontology population compared to single-agent baselines. In parallel, a reasoning layer combines OWL-DL with SWRL rules to derive patient-specific interpretations that reflect both medical and contextual nuances. We evaluate PATCHES across two dimensions: ontology quality (complexity, adaptability, and consistency) and ontology population (accuracy, robustness, and hallucination mitigation). Results demonstrate that our approach provides a scalable and explainable mechanism for contextual interpretation of ADHD, bridging the gap between theoretical models and practical application. This work offers a significant step toward adaptive therapeutic and educational strategies informed by context-aware, ontology-driven reasoning.

Item Type:	Article
Identification Number:	10.1016/j.eswa.2026.133339
Dates:	Date Event 1 June 2026 Accepted 18 June 2026 Published Online
Uncontrolled Keywords:	Contextualization, Ontology, SWRL rules, Multi-agent, LLM, ontology population, ADHD
Subjects:	CAH11 - computing > CAH11-01 - computing > CAH11-01-01 - computer science
Divisions:	Architecture, Built Environment, Computing and Engineering > Computer Science
Depositing User:	Gemma Tonks
Date Deposited:	13 Nov 2026 11:59
Last Modified:	13 Nov 2026 11:59
URI:	https://www.open-access.bcu.ac.uk/id/eprint/17310

Actions (login required)

View Item