A classification approach to interpretation of traditional Chinese musical score

Gongchepu, one of the widely-used traditional Chinese musical scores, is hard to read due to the immeasurable rhythmic rule, which only presents a general rhythmic structure while the duration of each note is determined by performer according to the context of the melody. Experience of determining the duration of each note is passed down via oral tradition, and there are few experts who can read such musical score now. However, by capitalizing on classification methods, such as Naive Bayes and Maximum Entropy Model, rhythms of gongchepu can be labeled and interpreted into staff automatically, making it much easier to read. A precision of 85.63% is achieved in experiments. As an attempt to solve the rhythmic immeasurability problem in the study of musical score with the application of statistical model, this work is conducive to the preservation of Chinese traditional cultural heritage.