Music Vocal Spectrum Classification Model based on Improved SVM Algorithm

Music vocal spectrum classification model based on improved SVM algorithm is studied. There are two main types of the current music classification and detection models in complex noise environmentsn manely the linear and non-linear methods. Through the in-depth analysis on comparing the differences, this paper select the improved SVM as the tool. We firstly use the preprocessing model to normalize the collected vocal signal to serve as the basic step for the later information processing; Then, we conduct the music signal characteristic analysis to be the basis for the feature selection; Lastly, we design the novel improved SVM algorithm to classify the signal. After simulating the proposed model with MATLAB, the performance is visualized.