CAJ | 학술논문

隐马尔科夫支持向量机(HMSVM)是一种新颖的结构化支持向量机模型,该模型在序列标注学习任务比如英文组块分析中的有效性已经被证明,将该模型用于中文浅层句法分析识别任务,实验结果表明,该模型获得了较好的准确率和召回率.
은마이과부지지향량궤(HMSVM)시일충신영적결구화지지향량궤모형,해모형재서렬표주학습임무비여영문조괴분석중적유효성이경피증명,장해모형용우중문천층구법분석식별임무,실험결과표명,해모형획득료교호적준학솔화소회솔.
Hidden Markov Support Vector Machines is a novel structural SVMs model. Its efficiency has been proved in label sequence learning task such as English text chunking. In this paper,we treat Chinese chunk recognition as a label sequence learning problem. After giving the definition of Chinese shallow parsing,we apply HMSVM to solve Chinese shallow parsing problem. The results of experiment show that it achieves a better precision and recall.