摘要
XMMT系统排歧的“约束”方法研究深度不足,排歧效果不理想,因此通过嵌入外挂函数改善其排歧效果,可以提高汉英机器翻译准确度。设计一种二元语义模式规则库,并于XMMT系统中嵌入了Semantic Value语义函数,计算待归约成分与二元语义模式库的匹配度进行排歧,使得优化后的系统排歧效果得到改善。在对2000组随机短语进行排歧时,词义排歧正确率为79.9%,结构排歧正确率为85.7%,比原系统分别提高了8.6%和3.9%。
In XMMT system disambiguation“constraint”method,research depth is insufficient,disambiguation effect is not ideal,hence,embedding plug-in function is used to improve its disambiguation effect,it can improve the accuracy of Chinese English machine translation.A binary semantic pattern rule base is designed,and the semantic value function is embedded in XMMT system to calculate the matching degree between the reduced components and the binary semantic pattern base for disambiguation,which improves the disambiguation effect of the optimized system.The correct rates of word sense disambiguation and structure disambiguation are 79.9%and 85.7%,respectively,which are 8.6%and 3.9%higher than the original system.
作者
吕洋
张静
华芳
LÜYang;ZHANG Jing;HUA Fang(Basic Subject Department of Modern College, Northwest University, Xi’an 710000, China)
出处
《微型电脑应用》
2021年第11期19-21,共3页
Microcomputer Applications
基金
陕西省教育厅2020年度人文社科专项计划科研项目(20JK0395)。
关键词
二元语义模式
语义评价函数
排歧
XMMT系统
binary semantic pattern
semantic evaluation function
disambiguation
XMMT system
