汉藏短语抽取.pdfVIP

汉藏短语抽取 诺明花1,2张立强1，刘汇丹1’2，吴健1，丁治明1 1．中国科学院软件研究所，北京100190 2．中国科学院研究生院，北京100190 E-mail：{minghua,zlq,huidan,wujian,zhiming}@iscaS．ac．cn 摘要：本文将从汉藏法律法规和公文领域平行语料中提取双语短语对．考虑现阶段藏文资源不足，提出两步汉 藏短语抽取方法。第—步是提取汉语语块，这部分工作不是本文工作重点。第二步是获取待翻译汉语短语的译文， 该模块提出藏文词序列相交算法抽取藏文短语。该算法可以很好的抽取1．I和1-n连续和非连续藏文短语． 关键词：汉藏短语抽取：藏文信息处理：中文信息处理 Chi IqeseTi betanPhraseExtracti on MinghuaNu01一，LiqiangZhan91，HuidanLiul，一，JianWul，ZhimingDin91 1．InstituteofSoRware，ChineseAcademyofSciences，Beijing，100190 2．Graduate ofthe UniversityChineseAcademyofSciences，Beijing,100190 E·mail：{minghua,zlq，huidan，wujian，zhiming}@iscas．ac．cn Abstract：ThisdescribesamethodtOextract from Chinese·Tibetan paper phrase domain-specific bilingual pairs irIlawsand andofficialdocuments．Soused extractionmethod corpus regulations far,widelyphrase heavily ontheresultof andSO word oradditionalresourceslike or depends alignment part-of-speechsyntacticanalysis accountof resourcesiIlTibetanat forth．Taking inadequate two·phaSe present,thispaperproposes Chinese—Tibetanextractionmethod．Firstis of extractionChinese chunk) phrasepairs step phrase(multi—word and Reduction secondisextractionofcandidate usingNagao’SAlgorithmSubstring step Algorithm．The Tibetantranslationfor Chinese Tibetanwords translation·readyphrase．Thispaperproposes sequence intersection