The performance analysis of support vector machine parameters based on statistical analysis

By using support vector machine learning method for audio classification, we compare the classification accuracy of different parameter combinations and analyze parameters impact on classification accuracy, number of support vectors and classification time. According to the statistical analysis, we find that the value of sigma² has more influence on classification accuracy, number of support vectors and classification time than that of C. Based on the analysis, we propose a new parameter selection method about how to improve classification accuracy and reduce classification time.