蒙古族人名译名的自动识别研究  

Research on Automatic Transliterated Mongolian Names Identification

作　　者：贺慧[1] 

机构地区：[1]内蒙古工业大学信息工程学院,呼和浩特010080

出　　处：《内蒙古工业大学学报（自然科学版）》2015年第3期214-217,共4页Journal of Inner Mongolia University of Technology：Natural Science Edition

基　　金：国家自然科学基金资助项目(批准号:61363052);内蒙古工业大学科学研究项目(ZS201005)

摘　　要：蒙古族人名译名的自动识别对汉语自动分词的研究具有重要意义。本文提出了一种在中文文本中自动识别蒙古族人名译名的算法。以蒙古族人名译名用字信息为基础,利用译名首尾字表、称谓词表和指界动词表对候选译名再定界,并针对译名特点采用译名音节串成词的优先策略。实验结果表明,召回率达到97.67%,正确率达到95.26%。Transliterated Mongolian names identification is significant to the approach of Chinese word segmentation.This paper presents an effective algorithm for automatically identifying transliterated Mongolian names in Chinese texts.Based on the information of character that transliterated Mongolian names used,the boundary of candidate transliterated name is defined through the first character table of transliterated Mongolian names,the last character table of transliterated Mongolian names,the appellation table and the specified bounds table again.To further improve the result,also does the priority word-building strategy of syllable string for transliterated names in terms of the feature of transliterated Mongolian names.The preliminary experiment shows that the recall rate reaches 97.67% and the precision rate reaches 95.26%.

关 键 词：蒙古族人名译名 译名自动识别 汉语分词 中文信息处理 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]