Abstract: A machine translation system based on Case-Based Reasoning is described. We lay emphasis on the core of the system: the case retrieval part. The new similarity rule improves the translation accuracy with high coverage. The abstract-case-included case base and the recursively abstract algorithm promote the system's efficiency.

摘要： 介绍了一个基于范例推理的机器翻译系统。对其核心部分范例搜索作了深入的分析。提出的新的相似度量准则使系统在保证翻译覆盖率的前提下提高了翻译的准确度，而含有抽象范例的范例库和递归抽象搜索算法提高了系统的时间效率。

关键词: 机器翻译, 基于范例推理, 覆盖率

CLC Number: 

• TP391.2