Machine Learning Based Classificationcomparison Of Music Genres

Authors

  • Swati A. Patil, K. Thirupathi Rao

Abstract

Unlike western countries having specific music genres like blues, jazz, rock, pop, etc., in India, we can find the diverse nature of the music which varies from region to region. Classification of genres in music helps the music industry for music recommendation to the user. Due to the dynamic nature of audio signals especially music signals it is a great challenge to classify the music to get the best accuracy. Marathi is the regional language of Maharashtra state in India. In the presented work the samples of Marathi songs genres like balgeet, bhavgeet, bhaktigeet, samargeet, prarthana, lokgeet, lavni, Natyasangeet, wedding song are collected from the Marathi music collection for classification where training and testing are performed. This collection of music is used as a Marathi music dataset. Features play a very crucial role in any classification algorithm. For this Mel Frequency cepstral coefficients (MFCC), Short term Fourier transform (STFT), chroma, bandwidth, zero-crossing rate (ZCR), roll-off, mel spectrogram features are taken into consideration. Next various four-tier classification algorithms like Support Vector Machine (SVM), Decision tree, Random forest, K Nearest neighbor (KNN), Logistic regression are used for genre classification. The overall performance of the system with a three-tier Deep Neural Network (DNN) as compared to all the other classifications is notable.

Published

2021-08-12

How to Cite

Swati A. Patil, K. Thirupathi Rao. (2021). Machine Learning Based Classificationcomparison Of Music Genres. Drugs and Cell Therapies in Hematology, 10(1), 1614–1624. Retrieved from http://www.dcth.org/index.php/journal/article/view/313

Issue

Section

Articles