International Research Journal of Education and Technology

Artificial intelligence has significantly influenced digital education platforms, especially in the development of intelligent tutoring systems. However, many existing systems mainly evaluate student performance through test scores and activity completion, without considering the emotional condition of the learner. Emotional states such as confusion, boredom, frustration, or engagement can strongly influence learning efficiency and knowledge retention. This research proposes an Emotion-Adaptive Learning System that integrates multimodal behavioral signals to recognize the emotional state of students during the learning process. The system collects data from multiple sources including facial expressions, voice patterns, and keyboard interaction behavior. A hybrid deep learning framework is used to analyze these signals, combining convolutional neural networks for visual emotion detection and sequential models for audio-based sentiment analysis. The detected emotional state is then processed by an adaptive learning engine that dynamically modifies the educational content, pace, or difficulty level of the lesson. Experimental evaluation demonstrates that incorporating emotional awareness into learning platforms can improve student engagement and comprehension compared to traditional adaptive systems. The proposed approach contributes to the development of more intelligent and personalized educational technologies by integrating emotional intelligence with artificial intelligence-based learning environments.