International Journal of Innovative Research in                 Electrical, Electronics, Instrumentation and Control Engineering

A monthly peer-reviewed journal

ISSN Online 2321-2004
ISSN Print 2321-5526

Since  2013

Abstract: People's lives are greatly impacted by music. Like-minded people come together via music, and it serves as the community's binding agent. The genres of music that different communities write or even just listen to can be used to identify them. Our research aims to develop a machine learning system that can predict music genres more accurately than the current approaches.We constructed many categorization models for this project and trained them using the Free Music Archive GTZAN dataset. All of these models' performances have been compared, and the results have been recorded in terms of prediction accuracy. A select few of these models are trained using both the mel-spectrograms and the audio characteristics of the songs, while a select few others are trained exclusively using the spectrograms of the songs. One of the models, a convolutional neural network, was shown to have the highest accuracy of all the models when only given the spectrograms as the dataset.

Keywords: Music, communities, Listen.

PDF | DOI: 10.17148/IJIREEICE.2022.10735

Open chat