Machine Learning based Classification and Recognition of Melakarta Ragas using Melakarta Chakra Analysis

This study dives into the complex realm of Indian classical music, where the basic melodic compositions known as ragas—made up of musical notes—hold enormous pride and significance. The research especially focuses on the Carnatic tradition and uses machine learning approaches, such as Support Vector Classifier (SVC) and Random Forest model (RF), to accomplish accuracies of 89.65% and 99.72%. The high accuracy of the RF model underscores its effectiveness in classification and identification of ragas. To have a large data set and for the raga identification of 72 melakarta ragas overfitting method is used. By using 80/20 model training and testing of data is used to get the highest accuracy. In order to bridge the gap between tradition and technology in the world of Carnatic classical music, the study's accurate results utilizing both SVC and RF model classifiers pave the way for a deeper knowledge of ragas and their classification.