Harmonizing Emotions and Music with Fuzzy Intelligence for Personalized Recommendations

This study presents a music recommendation system that generates personalized playlists based on real-time emotional states, leveraging fuzzy logic to address the subjective nature of emotions. The system integrates a Fuzzy Nearest-Mean classifier for music emotion classification and a Fuzzy Inference System for facial emotion recognition. By adapting Russell’s circumplex model of emotions, the proposed approach enhances content-based music recommendation by incorporating emotion-based features. The system is evaluated using diverse datasets, including the Radboud Faces Database, achieving an average accuracy of 87% in music emotion classification and 91% in facial emotion recognition, even under challenging conditions such as partial occlusion and varied lighting. Experimental results demonstrate that the approach effectively improves user engagement and emotional connection with music, offering intelligent, emotionally relevant recommendations. This research contributes to the field of music information retrieval by integrating real-time emotion-aware recommendations for better user experience.