Geographic Entity Knowledge Representation And Management Based On AtomSpace

With the rapid development of geographic information technology,the expression and management of geographic entity knowledge has become one of the effective ways to maximize the value of geographical knowledge.In order to efficiently express the intricate relationship of geographical knowledge and organize these relatively isolated geographical knowledge into an efficient,unified and interrelated geographical knowledge base,knowledge map has become an effective way to organize knowledge.AtomSpace,as a graph database of knowledge representation(KR),can be used to express,store and manage knowledge,deduce it,and express knowledge in the form of hypergraph.Based on the advantages of AtomSpace database and the research on the relationship between geographical objects,this paper discusses the construction method of geographic knowledge map in AtomSpace from the perspective of geography and computer science,which breaks through the limitations of traditional graph database.It focuses on the conceptual theory of geographic entity knowledge representation,the study of geographic entity data model,the representation and management of geographic entity knowledge,the construction of geographic knowledge map and other research topics.It expresses the complex relationship between geographical objects more vividly and matches human cognition better.The research of this paper mainly covers the following aspects and conclusions:(1)Content expressed by geographical entities.By summarizing and analyzing the concept of traditional geographic entity knowledge,this paper makes a clear definition and definition of the geographical entity studied.It is concluded that the geographical entity referred to in this study not only includes the basic geographical elements and characteristics,but also includes the complex geographical entity that can express certain semantics and knowledge.The content of a geographic entity revolves around the characteristics of its name,location,shape,attributes and complex relationships.(2)Data model of geographic entity knowledge representation.Starting from the basic problems of geography,firstly,this paper analyzes the elements of knowledge representation of geographic entity in detail,including the types,attributes,relations and so on.Secondly,based on the division of geographical entity of land use by different industries and the characteristics of AtomSpace technology,the conceptual model and logical model of geographical entity knowledge representation are proposed.This data model not only ensures the neatness and accuracy of the geographic entity knowledge,but also makes the geographic entity knowledge more clearly expressed and understood.(3)The expression and storage management of geographic entity knowledge.In this paper,the hypergraph data model based on AtomSpace is used to represent geographic entities and the distributed characteristics of AtomSpace graph database are used to store and manage geographic knowledge.The research shows that the knowledge representation and management of geographic entity based on AtomSpace is more flexible and extensible than the traditional knowledge graph representation,and the query performance is more efficient,which can support more complex geographic entity relationship modeling.Therefore,it can meet the storage and management requirements of geographic entity knowledge,which can help people better understand and manage geographic entity knowledge.(4)Construction and experimental verification of geographic knowledge map.The geographic entity knowledge map was constructed around the topic of land use geographic entity and knowledge representation in different industry sectors,and the knowledge representation ability of geographic knowledge map was evaluated and analyzed.The experimental results show that the geographic knowledge map expressed by AtomSpace graph database can better express the complexity and richness of geographic entity knowledge,and perform better in statistical performance index.