摘要
提出一种基于知网的中文短信文本词汇链抽取方法。根据知网的语义关系,利用相同语义类给出上下文词汇项信息,构造多条词汇链,表达短信文本的多条叙事线索,从中抽取富含短信文本信息的词汇链,表达短信文本的语义信息,采用词汇链的关键词集合进行文本分类。实验结果证明,该方法的抽取准确率较高,文本分类速度较快。
This paper proposes a extraction method of Chinese short message text lexical chain based on HowNet.Information of context lexical item are given using same semantic class according to HowNet semantic relations.By constructing lexical chains or each,and some strong lexical chains with high quality can be extracted to express the semantic information of short messaging text and the key words using lexical chain set for text classification.Experimental results show that the accuracy of lexical chain construction method is superior,and greatly increases the speed of short messaging text classification.
出处
《计算机工程》
CAS
CSCD
2012年第10期67-69,73,共4页
Computer Engineering
基金
河北省科技支撑计划基金资助项目(10213581)
淮安科技计划基金资助项目(HAG09061)
关键词
知网
短信文本
词汇链
分类
关键词
统计信息
HowNet
short message text
lexical chain
classification
keywords
statistical information
