Perancangan Confusion-Aware Adaptive User Interface Berbasis Kombinasi Ekspresi Wajah dan Pola Navigasi Pengguna Pada Website Edukasi untuk Anak Berkebutuhan Khusus

The advancement of information technology has driven the use of educational websites as interactive learning media. However, most educational websites are still designed with static interfaces that do not consider user characteristics, particularly children with special needs (CSN), potentially causing confusion and learning barriers. This study aims to design a Confusion-Aware Adaptive User Interface based on a combination of facial expressions and user navigation patterns on educational websites for CSN. The system was developed using the Prototyping method with a rule-based approach to determine interface adaptation. Facial expression detection utilizes a combination of MediaPipe and Facial Emotion Recognition (FER), while navigation patterns are analyzed based on the number of clicks, misclicks, and page duration. Functional testing results using Black Box Testing indicate that all system features operate according to design. Testing of the Confusion-Aware Engine across 8 scenarios demonstrated a 100% classification accuracy of user states (normal, anxious, confused) with consistent implementation of the adaptive UI. The system successfully identified students through school-based unique code validation, displayed dynamic material according to animal categories (land, sea, air), and provided a comprehensive summary of learning outcomes equipped with PDF and Excel export features. This study contributes to the development of inclusive learning technology through adaptive interfaces that are responsive to user conditions in real-time.