| 融合句法信息的双语词对齐方法研究 |
| |
| 引用本文： | 张贯虹.融合句法信息的双语词对齐方法研究[J].数字社区&智能家居,2014(3):1519-1523. |
| |
| 作者姓名： | 张贯虹 |
| |
| 作者单位： | 合肥学院网络与智能信息处理重点实验室,安徽合肥260601 |
| |
| 基金项目： | 安徽省教育厅自然科学基金一般项目（KJ20128147）;安徽省高等学校省级优秀青年人才基金项目（2012SQRL171）;安徽省自然科学基金青年项目（1308085QF115） |
| |
| 摘 要： | 词对齐研究是多文种信息处理工作的一项不容忽略的基础性研究内容。通过针对中文和蒙古文词对齐研究中存在的形态和词序不对称以及支撑词对齐研究的对齐语料缺乏问题，开展融合语言信息的中蒙混合词对齐模型构建方法研究。利用产生式词对齐结果以及中蒙两种语言的语言信息作为潜特征，建立高质量的融合语言信息的中蒙混合词对齐模型。实验结果证明，该文提出方法对于利用可比语料抽取对齐语料是可行的。
|
| 关 键 词： | 句子对齐 短语翻译 中文短语 |
A Fusion Word Alignment Model Constructing Method Based on Linguistic Information |
| |
| ZHANG Guan-hong.A Fusion Word Alignment Model Constructing Method Based on Linguistic Information[J].Digital Community & Smart Home,2014(3):1519-1523. |
| |
| Authors: | ZHANG Guan-hong |
| |
| Affiliation: | ZHANG Guan-hong (Hefei University Key Laboratory of Network and Intelligent Information Processing, Hefei 260601,China) |
| |
| Abstract: | Generative model and discriminative model are taken as two mainly approaches in word alignment field. To solve the problems of Chinese-Mongolian word alignment such as asymmetric morphological structure and word order, is presented in a discriminative framework using the result of generative word alignment and Chinese and Mongolian contextual information as la-tent variables for Chinese-Mongolian word alignment task. This work provides a new method and some theoretical support for performance improvement of word alignment of smaller scale corpus. |
| |
| Keywords: | Sentence Alignment Phrase Translation Probabilistic Chinese Scripts |
| 本文献已被 维普 等数据库收录！ |
