An Ensemble Approach to Raga Classification Using Deep and Traditional Machine Learning

This research combines tonic pitch identification with deep learning and traditional machine learning algorithms for the classification of three types of Carnatic Raga music. Using Mel-spectrogram representations of audio data, we conducted multiple experiments, including Convolutional Neural Networks (CNN), ResNet18 with transfer learning, XGBoost and Random Forest classifiers. The ensemble model with a majority voting method achieved the highest test accuracy of 75.4%.