A MACHINE LEARNING APPROACH TO MUSIC MOOD CLASSIFICATION FOR EMOTIONALLY INTELLIGENT PLAYLISTS

Abstract

Music is a fundamental aspect of human life and has the power to evoke various emotions and moods. Understanding and categorizing music based on its emotional content, also known as music mood classification, has become a prominent area of research in recent years. Analyzing the emotional aspect of music is crucial in applications such as personalized music recommendations, mood-based playlist generation, and emotion-aware music therapy. The traditional approach to music mood classification involved employing music experts to listen to each track and manually assign mood labels, such as happy, sad, calm, energetic, etc. This process was highly subjective and prone to inconsistencies due to individual biases. The annotated data would then be used to build handcrafted rule-based systems or simple statistical models to classify music into different mood categories. While these approaches provided some insights, they were limited in scalability, generalization, and accuracy. In addition, manual annotation is time-consuming, expensive, and lacks objectivity. Moreover, human listeners may not always agree on the emotional interpretation of a particular piece of music, leading to discrepancies in the labeled data. To overcome these challenges and enable large-scale mood analysis of music collections, there is a demand for automated and data-driven approaches. Machine learning techniques offer a promising solution to this problem by leveraging computational models to learn patterns and relationships from data, thus enabling the automatic classification of music based on its emotional content. Therefore, this project develops an emotion recognition-based music recommendation system, which performs the mood analysis first, and then recommend the music according to the detected mood of the users. The experiments on real data confirm that the proposed mood classification system can be integrated to any music recommendation engine.