摘要
歧义问题的描述和消除问题是制约计算语言学发展的瓶颈问题.将交叉熵引入计算语言学消岐领域.采用语句的真实语义作为交叉熵的训练集的先验信息,将机器翻译的语义作为测试集后验信息,计算两者的交叉熵,并以交叉熵指导对歧义的辨识和消除.实例表明,该方法简洁有效,易于计算机自适应实现,交叉熵不失为计算语言学消岐的一种较为有效的工具.
The description of ambiguity and the disambiguation are the urgent difficuilities of computer Linguistics. To solve these questions, the cross entropy was introduced into the field of disambiguation. The real meaning of the sentence was regarded as the priority information of the cross entropy and the meaning of the sentence which was translated by computer was looked as the succeedinformation. Then the cross entropy was caculated and was used to direct the description of ambiguity and the disambiguation. The experiment results have shown its efficiency in solving the disambiguation problems. The cross entropy is a tool of disambiguation for computer Linguistics.
出处
《数学的实践与认识》
CSCD
北大核心
2006年第3期267-273,共7页
Mathematics in Practice and Theory
关键词
计算语盲学
歧义
消岐
交叉熵
computer Linguistics
ambiguity
disambiguation
cross entropy
