A study on Chinese traditional opera

In this paper, we have presented a study on content-based classification and similarity analysis among 8 typical genres of Chinese traditional opera. Acoustic features are extracted and different classifiers are evaluated, the results show that the quadratic classifier works best, its average classification accuracy can achieve 82.4%. After the similarity analysis, some interesting conclusions are also reported. The results of similarity analysis and automatic classification experiments among these 8 typical genres are coincident with each other. They are also coincident with our general knowledge of Chinese traditional opera.