作 者:王东明 徐金安 陈钰枫 张玉洁 WANG Dongming, XU Jin'an, CHEN Yufeng, ZHANG Yujie (School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, China) 机构地区:北京交通学院计算机与信息技术大学,北京100044 出 处:《中文信息学报》2017年第5期84-90,共7页Journal of Chinese Information Processing 基 金:国家自然科学基金(61370130,61473294); 中央高校基本科研业务费专项资金(2017JBM033); 国家国际科技合作专项资助(2017DFA11350) 摘 要:命名实体的翻译等价对在跨语言信息处理中非常重要。传统抽取措施通常使用平行语料库或可比语料库,此类措施受到语料库资源的质量和规模的限制。在日汉翻译领域,一方面,双语资源相对匮乏;另一方面,关于汉字命名实体,通常使用汉字对照表;关于日语纯假名的命名实体,通常采用统计翻译模型,此类措施受到平行语料库的质量和规模的限制,且精度低下。针对此问题,该文提出了一种基于单语语料的面向日语假名的日汉人名翻译对自动抽取措施。该措施首先使用条件随机场模型,分别从日语和汉语语料库中抽取日语和汉语人名;然后,采用基于实例的归纳学习法自动获取人名实体的日汉音译规则库,并通过反馈学习来迭代重构音译规则库。使用音译规则库计算日汉人名实体之间的相似度,给定阈值判定人名实体翻译等价对。实验结果表明,提出的措施简单高效,在实现系统高精度的同时,克服了传统措施对双语资源的依赖性。Named entity translation equivalents play a critical role in cross-language information processing. The tra- ditional method is usually based on large-scale parallel or comparable corpus, which is limited by the size and quality of the corpus resources. In Japanese-Chinese translation, the bilingual corpora resources are relatively scarce: the Chinese Hanzi and Japanese Kanji mapping table is often adopted to deal with Chinese named entity and a SMT mod- el to deal with the Japanese named entities in pure kana. In this paper, we propose a monolingual corpora based ap- proach. Firstly, the conditional random field model is adopted to extract Japanese and Chinese names from monolin- gual corpus. Then the Japanese-Chinese transliteration rule base is developed by instance based inductive learning in a iterative process employing the feedback learning. Experimental results show that the proposed method is simple and efficient, leverging the severely dependency on bilingual resource by the classical methods. 关 键 词:音译 分 类 号:TP391[自动化与计算机技术—计算机运用技术;自动化与计算机技术—计算机科学与技术] ,日语毕业论文,日语毕业论文 |