We developed a multi-class classification model to determine the genre of novel songs using a database of approximately 18,000 songs described by 15 numerical features. Our approach involved using a Random Forest model and an SVM to compare their performance. The Random Forest classifier achieved an accuracy of around 54% in genre classification, while the Support Vector Machine achieved an accuracy of approximately 56%. Although certain songs with ambiguous characteristics posed challenges, our model outperformed random chance and highlighted the importance of more high-dimensional data for accurate music genre classification.
bvnarayanan / music-genre-classification_ml_report Goto Github PK
View Code? Open in Web Editor NEWWe built a music genre classification model using Random Forest and SVM. Accuracy: Random Forest ~54%, SVM ~56%.