基于知识单元的自然语言结构化解析模型——以城市规划领域规则为例

An analytical model for structuring natural language based on knowledge unit——taking domain rules of urban planning as examples

将自然语言描述的知识规则应用于计算机处理时,需要将非结构化语言转换为结构化语言。对于面向特定领域的、具有空间关系的规则,还需要消除两种语言之间的空间信息语义障碍。本文以城市规划领域的知识规则为例,提出知识单元概念、构建具有空间关系的知识单元体系结构、提炼面向领域的知识单元组合模式。在此基础上,建立自然语言解析模型,实现将自然语言描述的领域规则转化为知识单元组合模式,进而解析成面向GIS查询、分析、操作的结构化的计算机语言。

When processing knowledge rules described by natural language are used in computer,it is necessary to convert nonstructured language to structured one.Especially,for the particular domain rules with spatial information,to eliminate semantic obstacle of spatial information between the two kinds of language is a difficulty for transformation.Taking domain rules in urban planning as examples,a concept of knowledge unit was proposed in the paper,a knowledge unit system was constituted and the combination patterns of knowledge units were abstracted.Based on them,the authors constructed an analytical model for natural language.Applying this model,rules described by natural language could be structured into combination pattern based on knowledge units,subsequently,be transformed into structured computer language for GIS inquiry,analysis and manipulation.