摘要
介绍生成多语种天气预报文本系统中句子规划器的设计与实现.句子规划主要由两部分组成:句子优化和语言资源映射.前者按照一定的优化规则在句子之间消除冗余成分;后者根据具体语言特点对句子进行结构化处理,确定文本结构树并为树中各节点添加语言资源信息,以供表层生成处理时使用.
This paper introduces design and implementation of a sentence planner for multilingual generation system of weather forecasting. The sentence planner can be divided into two parts——sentence optimizing and language resource mapping. The former removes redundant information from the input structure so as to get the sentence content succinct and clear; the latter adapts the structure to specific grammar requirement of the target language and adds syntactic and lexical information to each component of the structure so as to derive a complete text structure tree. Focusing on multilingual generation, the paper makes an investigation on the design and implementation of a sentence planner.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
1998年第10期105-109,共5页
Journal of Shanghai Jiaotong University
基金
国家自然科学基金
上海市科委科技发展基金
德国大众基金
关键词
句子规划器
多语种生成
天气预报文本
人工智能
sentence planner
multilingual generation
text structure tree
sentence optimizing
sentence structure class
language resource mapping