MUSIC EMOTION DETECTION USING MULTI-LEVEL CLASSIFICATION

ABSTRACT Music is one of the most powerful forms of human expression, capable of conveying and evoking a wide range of emotions across different cultures, languages, and social backgrounds. With the rapid growth of digital music platforms, streaming services, and intelligent multimedia systems, there has been an increasing need for automated techniques that can accurately identify and classify emotions embedded within musical content. Music Emotion Detection (MED) has therefore emerged as an important research area at the intersection of machine learning, signal processing, artificial intelligence, and affective computing. The primary objective of music emotion detection is to analyze musical signals and determine the emotional state communicated by a piece of music, thereby enabling improved music recommendation systems, personalized user experiences, music therapy applications, and emotion-aware human-computer interaction systems. This study focuses on Music Emotion Detection using a Multi-Level Classification approach. Traditional emotion detection systems often employ a single-stage classification model that attempts to categorize emotions directly into predefined emotional classes. However, due to the complex and subjective nature of human emotions, such approaches may suffer from reduced accuracy and limited adaptability.