摘要
提出在面向数据的英汉机译系统中，组合基于实例的技术实现源语分析的一种方法·其中介绍了基于编辑距离测量检索近似实例，以及根据候选分析得分的评价机制，利用近似实例及其对应分析获取最终源语分析结果的实现过程·实验结果表明，源语分析的测试指标查全率与查准率都比较令人满意·
A method which implements source language parsing with combining example based technique in data oriented machine translation system was proposed. Its foundation is the knowledge source built successfully based on data oriented parsing technique. Two main processing procedures were presented. One is searching similar examples from example set based on measurement of edit distance. And the other is generating the final parsing results according to score evaluation mechanism of the candidates by using similar examples and their correspondent parses. The tested parameters parsing recall and precision are both satisfactory and the source language parsing is accomplished efficiently and accurately.
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
1999年第3期243-246,共4页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金
关键词
源语分析
面向数据
机器翻译
英汉机译系统
source language parsing, similar example, chain structure, edit distance, match expression.
