Obtain Better Accuracy Using Music Genre Classification Systemon GTZAN Dataset
J SuriyaPrakash, Sai Kiran
Abstract
The term “music genre” refers to any method of categorization that divides musical compositions (or even individual musical phrases) into distinct musical forms or styles. Religious music, secular music, classical music, and pop music are only few of the many musical genres that exist. The amount of data available to us is increasing rapidly, making it infeasible for manual curation. In this work, we apply a variety of simple and basic machine learning algorithms namely Logistic Regression, K-Nearest Neighbor, Random Forest, Support Vector Machine and Artificial Neural Network, along with dimensionality reduction techniques namely, PCA, KPCA and LDA, on the GTZAN dataset. Further, we compared their accuracies and found that the combination model of KNN with PCA provides the highest accuracy of 77.41% among the compared models.