ISCA Archive ISCSLP 2006
ISCA Archive ISCSLP 2006

Research and Analysis of Fast Training in SVM-based Audio Classification

Shilei Zhang, Hongchen Jiang, Shuwu Zhang, Bo Xu

In this paper, we propose a new method to choose the effective samples for support vector machines (SVM) training in audio classification task. The objective is to reduce the training time of SVM by choosing effective examples from the training set of binary classes. We test the performances of our new method on a dataset composed of about 6-hour audio data which illustrate that the computation time can be significantly reduced without a significant decrease in the prediction accuracy.