JIITA, vol.8 no.1, p.906-913, 2024, DOI: 10.22664/ISITA.2024.8.1.906
Alden Robell M. de Loyola, Robert R. Roxas
University of the Philippines Cebu, Philippines
Abstract: This paper presents a study on Philippine music genre classification. The dataset was manually created by sampling audio features from 1,400 Philippine music tracks on seven genres. After classifying the data, the models were evaluated using accuracy (model analysis) and recall (genre analysis). Findings show that k-nearest neighbors, support vector machine, and random forest were the best-performing models, while decision tree was the worst-performing model. Rondalla was the most predictable genre, followed by Kulintang, Kundiman, and Rap. Pop Ballad, Rock, and Manila Sound were the difficult genres to predict. This study implies that popular machine learning models work well with the classification of Philippine music.
Keywords: music genre classification, machine learning, Philippine music, audio features
Fullpaper: