摘要
提出并实现了一种基于支持向量机(SVM)的中文文本中人名的自动识别方法。对训练文本进行自动分词、词性标注及分类标注,然后按字抽取特征,并将其转化为二进制表示,在此基础上建立了训练集。然后通过对多项式Kernel函数的测试,得到了用支持向量机进行人名识别的机器学习模型。实验结果表明,所建立的SVM人名识别模型是有效的。
Based on the characteristics of person names in Chinese texts, a method of automatic recognition of Chinese person names using support vector machines (SVMs) is proposed. The character itself, character-based POS tag, the information whether a character appears in a last names table, the probability of a character's occurrence in person names and context information are extracted as the features of the vectors. Each sample is represented by a long binary vector, and thus a training set is established. The machine learning models of automatic identification of person names are obtained by testing polynomial Kernel functions. The results show that the models are efficient in identifying person names from Chinese texts. The recall, precision and F-measure are up to 92.14%, 96.43% and 94.24% respectively in open test.
出处
《计算机工程》
EI
CAS
CSCD
北大核心
2006年第19期188-190,201,共4页
Computer Engineering
基金
国家自然科学基金资助项目(60373095
60373096)
关键词
支持向量机
中文文本
人名识别
机器学习
Support vector machines(SVM)
Chinese texts
Recognition of person names
Machine learning