Evaluating Machine Learning Models for Musical Style Categorization

Listening to music is a universal and inclusive experience enjoyed across all age groups, offering manifold benefits such as mood enhancement, heightened mental alertness, memory improvement, and a sense of tranquility. The profound impact of music on emotions and memories can be attributed to the activation of the brain's limbic system. The release of dopamine, a neurotransmitter associated with happiness, has the potential to elicit powerful emotional responses, even inducing goosebumps. Rooted deeply in cultural traditions, music transcends mere style and form, encompassing genres like Pop, Hip-Hop, Rock, Jazz, Blues, Country, and Metal. This paper presents an automated music genre classification system designed to enhance users' exploration of music. Leveraging machine learning, specifically K-Nearest Neighbors (KNN), Random Forest, and Support Vector Machine (SVM), the study seeks to identify the algorithm that attains the highest accuracy in music genre classification. The analysis aims to surpass existing models by training various classification models on a dataset comprising 30-second audio files. Extracted features, including MFCC, tempo, harmony, roll-off, chroma, root mean square, and spectral data, are employed to assess and compare the performance of these models.