A Classification Model of Music Genres Using BP Neural Network Based on Genre Similarity Analysis
Dazhen Sun, Mo Lv, Hang Ren, Jiaqi Fan, Qifang Liu · 2021 China Automation Congress (CAC) · 2021
The classification of music is not only beneficial to art research but also conducive to the development of music. As one of the main research issues, the automatic classification of music genres has attracted many researchers' attention. To classify a genre into the category, this paper proposes a music classification method based on neural network by using a large number of music data as samples. Firstly, in order to quantify the similarity of music genres, hierarchical cluster analysis is carried out by using 14 indicators from characteristics of the music, type of vocals and music description. Then, the BP neural network music genre classification model is established based on the clustering analysis of music genres. Finally, the accuracy and effectiveness of the music genre classification model are verified through a data set of nearly 90,000 songs. The results show that the classification accuracy of the model can reach 91.4%.