摘要
当前中文词语的本体概念相似度计算主要是利用知网来进行计算,但这种方法不能及时更新,对于新出现的词语往往无法操作。为此,提出了一种自动实时更新的本体概念之间语义相似度计算方法,加入搜索引擎这个因素,利用知网的结构严谨和搜索引擎可以时时更新来提高算法的性能,以适应不断的信息爆炸式的发展,最后进行了实验分析比较。试验结果表明,该方法能比较好地处理一些新词或词组的新概念,符合人的主观判断,又有一定的严谨性。
Currently, semantic similarity computation of the concept in Chinese mainly used by HowNet, but this method can not be updated in time, so the new words are often inoperable. A new method of semantic similarity computation between concepts based on HowNet is presented, by adding a search engine of this factor. Search engine can improve the performance of algorithm to adapt to the explosive development of information. The experiments prove that the method can handle a number of new words or phrases of the new concept better. It is in line with people＇ s subjective judgments, and has a certain degree of rigor.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第19期4419-4421,共3页
Computer Engineering and Design
关键词
本体
语义相似度
搜索引擎
语义网
主观判断
ontology
semantic similarity
search engine
semantic web
subjective judgment
