【大学】双语语料库段落重组对齐方法研究

上传人:人*** 文档编号:568336248 上传时间:2024-07-24 格式:PPT 页数:32 大小:206KB
返回 下载 相关 举报
【大学】双语语料库段落重组对齐方法研究_第1页
第1页 / 共32页
【大学】双语语料库段落重组对齐方法研究_第2页
第2页 / 共32页
【大学】双语语料库段落重组对齐方法研究_第3页
第3页 / 共32页
【大学】双语语料库段落重组对齐方法研究_第4页
第4页 / 共32页
【大学】双语语料库段落重组对齐方法研究_第5页
第5页 / 共32页
点击查看更多>>
资源描述

《【大学】双语语料库段落重组对齐方法研究》由会员分享,可在线阅读,更多相关《【大学】双语语料库段落重组对齐方法研究(32页珍藏版)》请在金锄头文库上搜索。

1、双语语料库段落重组对齐方法研究 http:/ et al. (1991c)字符的个数-Gale and Church (1993)基于词典的句子对齐http:/ 句珠二分图(完全二分图)匹配 最优对齐匹配 http:/ bead) 每一对相对应的句子称作句珠http:/ V2V,V1V2,使G的每一条边eE时,e=vi,vj,viV1,vjV2,即同一子集Vi(il,2)中的任何两个结点都不邻接,称这样的图为二分图。http:/ http:/ 设G是二分图,如E的一个子集M中的边无公共端点,即任两边均不邻接,则M为G的一个匹配。最优对齐匹配http:/ ,满足以下条件:二分图中所有的节点是有序

2、的 若|S|=m,|T|=n则首先默认sm,tnE;M中若存在一条边e=si,tj满足d(si,tj) D(D为特定的阈值);此时,M中不存在边sk,tr使得kj或ki且rj成立;采用贪心算法依次在完全二分图G中选取权值最小的边,直到M中边数达到饱和。http:/ S2, S3, S4, S5, S6, S7 Sm-2, Sm-1, SmT1, T2, T3, T4, T5, T6, T7 Tn-2, Tn-1, Tn 羽嘱傅筋训甭盲颈甥拓东敖请戎立邢频涌催铅迫撕桔剁枯蕴刊侮堰角顷款【大学】双语语料库段落重组对齐方法研究【大学】双语语料库段落重组对齐方法研究段落对齐问题和模型的关系S(s1,

3、s2, s3,si, sj, sm) V1T(t1, t2, t3,ti, tj, tn) V2(1:1)型句珠(si, tj)的集合Km,n默认文本中最后一句对齐sm,tnE 双语段落重组对齐 寻找最优对齐匹配http:/ (1:1)型正确对齐句珠在文本中的分布情况 (1:1)的句对在全篇的比例均超过85%,并且有着良好的分布规律 http:/ Pi,j = a(Pui,j-P0) + (Pli,j -P0) + a(Pdi,j -P0)P0为对应文本长度之比;Pui,j为对应上文部分长度之比;Pli,j为对应句长度之比;Pdi,j对应下文部分长度之比;a为加权系数http:/ 汉语文本P0

4、锚点句对Pui,jPli,jPdi,j践磅叁凉戒梢收棋蕴慧精竣牙游炬届粗狸泡致瘁鸦赶站挂皇与鲁掐不贷迂【大学】双语语料库段落重组对齐方法研究【大学】双语语料库段落重组对齐方法研究锚点句对的校验词典检验 L| |表示全部元素的字符长度和;Match(S)表示译文出现在汉语句中英语单词;Match(T)表示成为英文单词译文的汉语单词 ;http:/ = (返回段落总数 错误段落总数)/ 双语文本中(1:1)型正确句珠总数;准确率 = (返回段落总数 错误段落总数)/ 返回段落总数;http:/ 重组分段的结果一般每一段落的句子数在10句以内。叹舆琅笑钵杯蛙费胀逃考协腹荆支怂不堡输槽她鹿林碴抗惧辫嗜

5、蔫庐标逮【大学】双语语料库段落重组对齐方法研究【大学】双语语料库段落重组对齐方法研究重组分段的实例(1)-originalAround 30 people have staged a naked protest against GM food. The protesters spelled out no GM with their bodies in a meadow at Forest Row, East Sussex.Organiser Mike Grenville said he hoped it would send a message to the Government of peo

6、ples concern, particularly over commercial planting of GM crops.Mr Grenville, 51, a self-employed business consultant from Forest Row, said: I think people were very pleased to have the opportunity to express how we feel, how frustrated we are about what seems to be the foregone conclusion, and the

7、question many people are saying is What do we do next?. We hope others will follow our lead and find other ways to express how we feel. We do not want GM crops planted in the country at all.http:/ 30 people have staged a naked protest against GM food. The protesters spelled out no GM with their bodi

8、es in a meadow at Forest Row, East Sussex.Organiser Mike Grenville said he hoped it would send a message to the Government of peoples concern, particularly over commercial planting of GM crops.Mr Grenville, 51, a self-employed business consultant from Forest Row, said: I think people were very pleas

9、ed to have the opportunity to express how we feel, how frustrated we are about what seems to be the foregone conclusion, and the question many people are saying is What do we do next?. We hope others will follow our lead and find other ways to express how we feel. We do not want GM crops planted in

10、the country at all.n约30名人士在英国苏塞克斯郡东部的福利斯劳的一处牧场裸体组成“反对基因改良作物”的字样。组织者迈克-格伦威利,希望通过这次行动向政府传达人们反对商业化种植基因改良作物的意愿。51岁的格伦威利是福利斯劳一名自雇的商业顾问,他说:“当人们得知种植基因改良作物似乎不可避免时,他们感到十分沮丧。许多人都提出这样的问题,接下来我们该做什么?。令人欣慰的是,人们有机会表达内心的感受。希望其他人加入我们,跟我们一起找到其他表达内心感受的方法。我们根本不希望在这个国家里种植基因改良作物。铭爬黍第九徽商苛弊亚赚吞匪叔透仁恶满凸徊炭抽垫香垦预抵旅抚档盖蝇【大学】双语语料库段

11、落重组对齐方法研究【大学】双语语料库段落重组对齐方法研究重组分段的实例(3)-ReAlignhttp:/ 30 people have staged a naked protest against GM food. The protesters spelled out “no GM” with their bodies in a meadow at Forest Row, East Sussex.Organiser Mike Grenville said he hoped it would send a message to the Government of peoples concern,

12、 particularly over commercial planting of GM crops.nMr Grenville, 51, a self-employed business consultant from Forest Row, said: I think people were very pleased to have the opportunity to express how we feel, how frustrated we are about what seems to be the foregone conclusion, and the question man

13、y people are saying is What do we do next?. nWe hope others will follow our lead and find other ways to express how we feel. nWe do not want GM crops planted in the country at all.n约30名人士在英国苏塞克斯郡东部的福利斯劳的一处牧场裸体组成“反对基因改良作物”的字样。组织者迈克-格伦威利,希望通过这次行动向政府传达人们反对商业化种植基因改良作物的意愿。n51岁的格伦威利是福利斯劳一名自雇的商业顾问,他说:“当人们得知种植基因改良作物似乎不可避免时,他们感到十分沮丧。许多人都提出这样的问题,接下来我们该做什么?。令人欣慰的是,人们有机会表达内心的感受。希望其他人加入我们,n跟我们一起找到其他表达内心感受的方法。n我们根本不希望在这个国家里种植基因改良作物。勿筷吧乳嫁肃秉剿戍降阶假卡瓦处纸聋坠健铁辕屯译椒歪藕策铣烹励所史【大学】双语语料库段落重组对齐方法研究【大学】双语语料库段落重组对齐方法研究主要内容引言方法结果讨论http:/

展开阅读全文
相关资源
正为您匹配相似的精品文档
相关搜索

最新文档


当前位置:首页 > 建筑/环境 > 施工组织

电脑版 |金锄头文库版权所有
经营许可证:蜀ICP备13022795号 | 川公网安备 51140202000112号