摘要
介绍汉语天气预报自动生成系统中句子结构优化的方法。该系统通过分析气象数据而生成天气预报文本，文本在计算机内部使用文本结构树表示，该结构通常具有一定的冗余性。采用一种基于语义的规则驱动方法来消除句子之间的信息冗余，从而使生成的文本更加自然流畅，增加可读性。句子优化的主要类型、系统的构造以及规则的定义方法均在文中作了详细的说明。
In this paper we address the problem of sentence structure optimizing in the sentence planner of the system MLWFA which aims at generating fluent and correct Chinese weather reports. We adopt a semantic-based rule driven method to cut down redundancy in the text structure tree, which is quite unique and appropriate for natural language generation system. This paper identifies various types of optimizing, and sets forth the rule definition mechanism as well as some constraints. A detailed example is also shown here.
出处
《计算机工程》
CAS
CSCD
北大核心
1998年第7期14-16,58,共4页
Computer Engineering
基金
国家自然科学基金!69673008
上海市科委科技发展基金!962907002
德国大众基金
关键词
句子结构优化
自然语言生成
信息处理
天气预报
Sentence structure optimizing
Chinese text generation, Semantic-driven rules
Text structure tree
