基于句子相似古今句子检索.pdfVIP

基于句子相似的古今句子检索* 康生巧，白宇，韩亚冬，蔡东风 沈阳航空航天大学知识工程中心,辽宁沈阳110136 E-mail: bakangs@126.com 摘 要：古籍文献中蕴含着丰富的知识，如‘医古文’在现代中医理论发展过程中仍占据重要基础地位， 因此面向古籍文献的检索技术研究具有一定的应用价值。然而，目前文本信息检索技术的研究主要面向由 现代汉语构成的文本，针对古汉语文献资料的检索研究并不多见。本文提出一种基于共现字序列的古今句 子相似度计算方法，实现了古文与现代文句子之间的跨文种检索。利用本文方法在《内经》原文与现代文 对照版本上进行了古文与现代文的双向句子检索实验，实验结果表明本文方法在古今句子检索上取得较好 效果。 关键词：信息检索；二元文法；古文检索； Sentence retrieval between ancient and modern Chinese based on sentence similarity KANG Sheng-Qiao, BAI Yu,HAN Ya-Dong, CAI Dong-feng ( Knowledge Engineering Research Center , Shenyang Aerospace University , Shenyang , Liaoning 110136 , China) E-mail: bakangs@126.com Abstract: Ancient books written by ancient Chinese contain a wealth of knowledge, for example, ancient Chinese books concerning medicine had played a fundamental role in the development of modern theory of Traditional Chinese Medicine. Therefore, the study of retrieval technology focusing on ancient books is proved to be of practical value. However, current research on text information retrieval technology mainly focuses on the texts that are composed of modern Chinese, the research on retrieval technology of ancient books are rarely seen. This paper brings forward a calculation method of Similarity between the ancient and modern Chinese based on the sequence of words that are commonly used in the two versions so as to realize cross-language retrieval between ancient and modern Chinese. Experiments on the ancient Chinese book “Canon” and it modern Chinese version show that the proposed algorithm achieve encouraging results. Keywords: information retrieval ; bi-gram ; ancient text retrieval 1 引言 “医古文”是中国古人留给后人的宝贵遗产，它记载着宝贵的医学理论和丰富的医疗 经验，表述具有