基于卡方检验的汉语术语抽取_胡文敏.pdfVIP

第 27卷第 12期 计算机应用 V ol. 27 No. 12 2007年 12月 ComputerApplications Dec. 2007 : 1001- 9081( 2007) 12- 3019- 02 1, 2 1, 2 1, 2 胡文敏 , 何婷婷 , 张 勇 ( 1. 华中师范大学计算机科学系, 武汉 430049; 2. 国家语言资源监测与研究中 网络媒体分中 , 武汉 430049) (naiheye6829@ hotm ai.l com) : 发现术语在中文信息处理和语言学习方面具有非常重要的作用和意义提出了一种基 于卡方检验的汉语术语抽取方法: 先从网络上下载语料, 然后使用改进的互信息参数( -M I)抽取结 构简单的质串, 并在其基础上进一步使用卡方检验结合质子串分解方法抽取具有复杂结构的合串 实验结果显示, 该算法有效地提高了汉语术语抽取的精确度 : 卡方检验; 质子串分解; 互信息 : TP391. 43 : A Extraction of Ch inese term based on Ch i- square test 1, 2 1, 2 1, 2 HU W en-m in , HE T ing-ting , ZHANG Yong ( 1. Department of ComputerScience, H uazhongN ormal University, WuhanH ubei 430079, China; 2. onitor and Research Centerf orNationalLanguageResourceN ewt ork ultmi edia Sub-branch Center, WuhanHubei 430079, China) A bstract: D iscovering the term has very mi portant app lications in Chinese inform ation processing and language learn ing. A m ethod for the extraction of Chinese term based on Ch-i square testw as proposed. irst, dow nloadW eb docum ents and bu ild a corpus, then prmi ew ordsw ere extracted by using the -M I param eter mi proved by mutual-information, wh ile com b inedw ords were extracted by the Ch-i square testw ith the help of decom position of prmi e string. The expermi ents show that the algorithm can effectively mi prove the precision in the extraction of Chinese term. K ey w ords: Chi-square test; decom position of prmi e string; mutual inform ation [ 3, 4] 0 引言 ,