摘要
用支持向量机(SVM)方法对中文地名的自动识别进行了探讨,对于含特征词的地名和非地名用支持向量机进行分类:结合中文地名的特点,抽取地名构词可信度及其前后词的词性作为特征向量的属性,建立了一定规模的训练集,并通过对不同kernel函数的测试,得到了地名分类的机器学习模型.实验表明,对于切分正确的地名,本方法具有良好的效果.
This paper presents a method of automatic recognition of Chinese place names based on support vector machines （SVM）, the place names and non-place names are classified by SVM： considered the feature of Chinese place names, the lexical reliability of place names and part-of-speech around the place names are extracted as the attributes of feature vectors. A training set is established. The machine learning models of automatic classification of Chinese place names based on support vector machines are obtained hy testing different kernel functions. The results of some preliminary tests show that the SVM method is efficient for classifying Chinese place names.
出处
《小型微型计算机系统》
CSCD
北大核心
2005年第8期1416-1419,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金(60373095
60373096)资助
关键词
支持向量机
中文地名识别
机器学习
support vector machines
recognition of Chinese place names
machine learning
