ARCHIVES
VOL. 10, ISSUE 4 (2025)
Performance evaluation of selected machine learning models for music genre classification
Authors
Oguntola KM, Olabiyisi SO, Ismaila WO, Saka AO, Ogunleye TO
Abstract
The growing dominance of digital streaming has
made efficient music classification models essential, particularly for Nigerian
music, which reflects diverse cultural expressions and linguistic variations
that do not always align with Western genre systems. This study investigates
three machine learning techniques—K-Nearest Neighbors
(KNN), Support Vector Machine (SVM), and Decision Trees (DT)—to determine
their effectiveness in genre classification. Using Librosa, Mel-Frequency
Cepstral Coefficients (MFCCs) were extracted from audio samples to represent
musical features. These features were then applied to the models: KNN, which
classifies based on similarity; SVM, which identifies an optimal boundary for
separation; and DT, which organizes decisions hierarchically. The models were
evaluated using accuracy, precision, recall, and F1-score. Results revealed
that SVM achieved superior performance with 93% accuracy, while KNN and DT
followed with 88% and 79% respectively. These findings confirm SVM as the most
reliable model for music classification, offering strong potential for
enhancing intelligent music retrieval and personalized digital libraries.
Download
Pages:11-17
How to cite this article:
Oguntola KM, Olabiyisi SO, Ismaila WO, Saka AO, Ogunleye TO "Performance evaluation of selected machine learning models for music genre classification". International Journal of Advanced Scientific Research, Vol 10, Issue 4, 2025, Pages 11-17
Download Author Certificate
Please enter the email address corresponding to this article submission to download your certificate.