Chinese dialect identification using clustered support vector machine
Mingliang Gu, Yuguo Xia · 2008
This paper presents a novel Chinese dialect identification method to solve the poor decision ability existed in most dialect identification system. The new method firstly uses Gaussian mixture models and n-gram language models to produce a global language feature, and makes decision using clustered support vector machine. The experimental results show that the new method not only raises correct identification rate greatly, but also improves the robust of the system.