节点文献
基于地理特征类型的自然语言空间关系查询
Natural-language Querries on Spatial Relations Based on Geographic Entities
【作者】 司望利；
【导师】 郭新成；
【作者基本信息】 长安大学 ， 地图制图学与地理信息工程， 2008， 硕士
【摘要】 随着地理信息系统越来越深入应用于人们的日常生活之中,面向自然语言的空间关系查询研究越来越被人们所重视,已成为一个研究热点。目前在面向自然语言的空间关系查询方面的研究主要集中在空间关系描述、空间关系自然语言形式化表达、空间关系自然语言查询语句转换以及空间关系自然语言查询接口等方面,但在研究时通常将参与的空间对象抽象为不具有地理语义的空间几何体,很少考虑在空间关系自然语言表达时,不同的地理特征类型所引起的语义差别影响,因此在具体的自然语言查询实现时还难以完成面向现实中客观存在的空间物体的查询要求。本文首先依据目前在空间关系的自然语言描述词汇形式化表达方面的研究成果,利用已有的形式化表达规则,设计出具体定性的自然语言描述词汇在计算机模型中所对应的定量化空间关系判断算子,以此作为空间关系自然语言查询时空间关系选定的依据。同时,针对不同特征类型的空间地物和空间关系自然语言描述词汇之间对应的语义关系有所不同的特点,文章在收集总结自然语言词汇在描述空间关系时的对应知识的基础之上,利用产生式知识规则表达的方法,提出不同特征类型地物和描述词汇间的知识规则库的设计方法。最后通过部分空间地物的规则实例,结合具有GIS功能的平台,设计出具体查询系统实践举例证明。
【Abstract】 With more and more applications of geographic information system in people’s daily life, study on the natural-language queries of spatial relations becomes increasingly important.Currently, Scholars focus on natural-language descriptions of spatial relations, formalizing descriptions of natural-language spatial relations, parsing the query sentences , and the interface of natural-language spatial relation query, but they usually abstractly consider the spatial objects as non-geographic geometric objects when they dealing with their spatial relations. However, geographic semantics also take influences when describing spatial relations with natural language due to the characteristics of different geographic entities, so current study on the natural-language queries of spatial relations is hard to get a sound result when it faces to real world geographic objects.In this dissertation, firstly, the operators about the natural-language terms which describe spatial relations are designed according to the rules from former study on formalizing natural-language spatial relations descriptions. At the same time, considering the semantics of geographic objects, this paper sums up some knowledge about how the spatial relations between two different geographic objects which have different geographic characteristics are represented in natural language, and puts forward a method to design a knowledge base. Finally, use the rules in knowledge base and GIS platform to design a query system.