汉语NP识别方法的实验比较研究

The Test Comparison Research on Chinese NP Recognition Methods

利用错误驱动法、支持向量机法和隐马尔可模型3种方法对汉语文本进行名词短语识别,对实验进行比较分析,结果表明SVM与HMM的识别效果总体上要好于错误驱动法,HMM法在封闭测试中优势明显.研究表明错误驱动法适用于解决从语料库中学习转换规则的传统问题;SVM方法适用于解决两类别的分类问题;而HMM方法侧重应用在与线性序列相关的现象上.

This paper adopts three methods including Error-Driven, Support Vector Machines and Hidden Markov Model to recognize noun phrases in Chinese texts. Through the comparison and analysis to the experiments , the result shows that the recognition effect of the latter two methods is better then the effect of the first one, and the function of HMM in the closed test is dominant. The study shows that the Error-Driven method is often used to solve the traditional problems of learning transformation regulations from corpus, SVM is adaptable to solve the classification problem of two categories, and that HMM is mainly used on the problems related to linear array.