CAJ | 학술논문

本文在中文维基百科知识库的基础上，对文本语义相关度计算进行了研究。实验选取了2014年12月15日在中文维基百科网站下载的主题文章，进行处理后作为语义概念知识库。在Words－240测试集上的实验结果表明，该方法比基于WordNet的LSA算法的效果要好。
본문재중문유기백과지식고적기출상，대문본어의상관도계산진행료연구。실험선취료2014년12월15일재중문유기백과망참하재적주제문장，진행처리후작위어의개념지식고。재Words－240측시집상적실험결과표명，해방법비기우WordNet적LSA산법적효과요호。
In text semantic understanding , massive quantity of common sense and specialized knowledge is needed;it doesn ’ t suffice to only use human-compiled dictionary and thesaurus .As social networks develop , on-line Wikipedia provides a platform for sharing and improving human knowledge .This paper , on the basis of Chi-nese Wikipedia , studies the calculation of text semantic relativity .A corpus of processed texts is used as knowledge base of concepts;the texts are downloaded from Chinese Wikipedia as of 2014 December 15 .The results of experi-ment on Words-240 test set indicate that the method discussed in this paper is superior to WordNet-based approa-ches and LSA method .