HMM和负反馈模型在词性标注中的应用  

The application of HMM and negative feedback model in POS tagging

作　　者：温锐[1] 朱巧明[1] 李培峰[1] 

机构地区：[1]苏州大学计算机科学与技术学院,江苏苏州215006

出　　处：《苏州大学学报（自然科学版）》2005年第3期39-42,共4页Journal of Soochow University(Natural Science Edition)

基　　金：江苏省自然科学基金资助项目(BK2003030);江苏省高校自然科学基金资助项目(04KKB320134)

摘　　要：词性标注是词法分析的基础.HMM是一个数学模型,具有算法成熟、效率高、易于训练的优点;负反馈是一个控制模型,在规则中引入负反馈,可以增强规则的客观性.本文探讨了两个模型,然后将它们应用于词性标注中.实验结果表明两者的结合是有效的和成功的.Part of speech tagging is the basis of lexical analysis. HMM is a mathematical model, which has mature algorithm, high efficiency and is easy to train; negative feedback is a control model, which can strengthen the objectivity of the rules when it is introduced into them. The two models are discussed and applied to part of speech tagging. The result of experiment shows that the combination of them is effective and successful.

关 键 词：词性标注 HMM 负反馈 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]