基于音频特征的SVM音乐风格识别  

Music styles recognition using SVM based on audio features

作　　者：张珺[1] Zhang Jun(Shangluo Vocational and Technical College,Shangluo,726000,China)

机构地区：[1]商洛职业技术学院,陕西商洛726000

出　　处：《现代科学仪器》2019年第3期44-47,共4页Modern Scientific Instruments

摘　　要：为提高音乐风格识别的准确率,提出了一种基于样本熵和SVM的音乐风格识别模型。通过EMD分解提取不同音乐风格信号的样本熵,将样本熵作为SVM的输入,音乐风格类型作为SVM的输出,建立SVM的音乐风格识别模型。研究结果表明,与RBFNN、BPNN和KNN相比,SVM音乐风格识别准确率高达98.43%,从而说明SVM可以有效提高音乐风格识别的准确率。To improve the accuracy of music style recognition,a model based on sample entropy and SVM is proposed.The sample entropy of different music style signals is extracted by EMD decomposition,the sample entropy is used as the input of SVM,and the music style type is used as the output of SVM to establish SVM's music style recognition model.The results of the study show that compared with RBFNN,BPNN and KNN,SVM music style recognition accuracy rate is as high as 98.43%,which shows that SVM can effectively improve the accuracy of music style recognition.

关 键 词：音频特征 支持向量机 经验模态分解 音乐风格 样本熵 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]