Automatic Music Genre Classification Based on Linguistic Frequencies Using Machine Learning

• Published in: 2021 International Conference on Recent Advances in Mathematics and Informatics (ICRAMI) pp. 1 - 5
• Main Authors: Rao, Mekala Srinivasa, Pavan Kalyan, O., Kumar, N. Naresh, Tasleem Tabassum, Md, Srihari, B.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 21.09.2021
• Subjects: automatic classification; Classification algorithms; Convolutional neural networks; Feature extraction; genre; genre classifier; linguistic content; Linguistics; Music; recommender system; Time-frequency analysis; Visualization
• DOI: 10.1109/ICRAMI52622.2021.9585937

Classifying various music into its genre has a lot of applications in the real world. It plays an important role in several online music streaming services such as Gaana, Spotify etc. Most of the music recommender systems implement such feature. Over the past two decades music coming from various sources has been increasing at a high speed. Several musical communities are emerged based on the music genre. Therefore, in order to satisfy their requirements, the need for an automatic music genre classifier became evident. In the process of determining the genre of a music, accuracy of the prediction must be well maintained. In our project we are automatically classifying an unknown music into its genre with an effective accuracy. We are separating the linguistic content from the noise while extracting features from the set of audio files. This helps in obtaining a good accuracy of prediction. We are implementing various Machine Learning Algorithms to build our project. We considered the GTZAN dataset [4], which contains 1000 music files of 10 different genres with each file having a duration of 30 sec.