Music Style Recognition System Based on Support Vector Machine Algorithm
Music style classification is an important part of music information retrieval and music recommenda-tion.It requires higher and higher efficiency of music style classification.However,the recognition of music style is relatively difficult for non-professionals,so a music style recognition system has been established based on machine learning.Taking the four most representative music styles as its research object,this paper selected 47 jazz style music,47 rock style music,42 classical style music and 40 modern style music as samples,applied fast Fourier transform,image embedding,and combined with the machine learning model to establish a music style recognition system based on the support vector machine algorithm,and finally realized the simultaneous recognition of four music styles.The average value of the AUC(area under the working characteristic curve of the subject)used by this model for blind testing was 0.871,and the classification accuracy rate was 71.7%.
Support vector machine algorithmmusic stylefast Fourier transform
NETL
NSTL
万方数据
