This Author published in this journals
All Journal IJHCS
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Music Genre Classification Using 1D Convolution Neural Network Falola, Peace Busola; Akinola, Solomon Olalekan
International Journal of Human Computing Studies Vol. 3 No. 6 (2021): IJHCS
Publisher : Research Parks Publishing LLC

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31149/ijhcs.v3i6.2108

Abstract

Music genre classification system is a system that is important to the users for effectiveness in the digital music industry. One of the effective ways of genre classification is in music recommendation and access to users. With accurate classification system built, songs can be readily accessed by the users when the genre of the song is known and recommendation of songs to the users is made easy. Also, automatic classification of genre is important to solve problems such as tracking down related songs, discovering societies that will like specific songs and also for survey purposes. In recent times, deep learning techniques have proven to be effective in several classification tasks including music genre classification. This paper therefore examines the application of 1D Convolutional Neural Network for music genre classification. A new dataset consisting of 1000 Nigerian traditional songs with seven genres was used for this work. As features extraction is crucial to audio analysis, seven low level features also known as content based features were extracted from the songs in the dataset which served as input into the classifier. Our results showed that the accuracy level of the system is 92.5% with a precision of 92.7%, recall of 92.5% and f1 score of 92.5%.