Autism Spectrum Disorder Detection Using Case-Based Reasoning: A Clinical Decision Support Framework

Abstract

Early detection of Autism Spectrum Disorder (ASD) remains critical for improving developmental outcomes, yet traditional diagnostic approaches face challenges of subjectivity, time-intensive assessments, and limited access to specialist expertise. This article presents a comprehensive Case-Based Reasoning (CBR) framework for ASD detection that imitates clinician expertise by retrieving and adapting solutions from historical cases. Our system structures patient data, incorporating developmental history, behavioral symptoms, and standardized screening scores into a reusable case base. The framework implements a CBR cycle: case representation using description logic, similarity-based retrieval via weighted Euclidean distance, rule-enhanced adaptation, and continuous case base expansion through validated retention policies. The system provides explainable recommendations by referencing specific historical cases, enabling clinicians to understand and validate diagnostic suggestions.  This study proposes a practical and transparent AI-assisted autism screening approach that complements clinical expertise.