Title: Automated Web Accessibility Evaluation Using Machine Learning Techniques Under WCAG 2.1 Guidelines
Authors: Abdurr Rahman
Volume: 9
Issue: 12
Pages: 26-32
Publication Date: 2025/12/28
Abstract:
Web accessibility is essential to ensure inclusive access to digital content for users with diverse abilities. Despite the availability of established standards such as the Web Content Accessibility Guidelines (WCAG) 2.1, many websites continue to exhibit significant accessibility violations, particularly in critical domains such as education and healthcare. Existing automated accessibility evaluation tools primarily rely on rule-based techniques, which are limited in detecting semantic and context-dependent issues. This paper proposes an automated web accessibility evaluation framework using machine learning techniques to assess WCAG 2.1 compliance. The framework extracts structural, visual, and semantic features from web pages and applies supervised learning models to classify accessibility violations across the four WCAG principles: Perceivable, Operable, Understandable, and Robust. Experimental evaluation demonstrates that ensemble-based machine learning models outperform traditional linear classifiers in detecting accessibility issues. The results indicate that machine learning-driven automation can significantly enhance large-scale accessibility assessment and support more inclusive web development practices.