An Intelligent ICT Framework for Emotion Classification and Adaptive Feedback in Psychological Education

Emotion-Aware Adaptive Learning in Psychology Education

Abstract

To enhance the effectiveness of psychological education, this study proposes an Intelligent ICT framework for emotion classification and adaptive feedback. The framework employs a hybrid deep learning architecture integrating Convolutional Neural Networks (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) networks to analyze multimodal inputs, including textual sentiment, vocal tone, and facial expressions, for accurate emotion recognition. Based on real-time emotional state detection, the system delivers personalized learning interventions and adaptive feedback to improve learner engagement and emotional well-being. Experimental evaluation conducted with 320 participants demonstrates an emotion recognition accuracy of 94.6% and an F1-score of 92.8%. Additionally, the proposed approach achieves a 23% improvement in learners’ psychological flexibility compared to conventional feedback systems. The results indicate that the proposed ICT framework effectively supports emotion-aware digital learning environments and enhances outcomes in psychological education.