Research On The Construction Method Of Knowledge Graph For Administrative Geographical Names

Administrative division is the abbreviation of administrative district dividing,and it is an effective means for the country to better manage the economy,politics and culture for various regions.The information of administrative geographical names record the past and present life of the city,especially for the past 70 years since the founding of since the People’s Republic of China,it is an exact reflection of China’s economy,history and culture,as well as the embodiment of China’s politics,policies and management,which has great value and research significance.Since the founding of the People’s Republic of China in 1949,the amount of information on changes in administrative divisions in China has been very large,but there is less deep use of this information,especially that,the development and research on economy,society and culture of various provinces and cities is in urgent need of support from information on changes of administrative division and spatial data.With the rapid development of artificial intelligence technology in the information age,especially due to the effectiveness of knowledge graph technology on the analysis and expression of multi-source,heterogeneous and other massive data,knowledge graph technology provides a new way and idea for the processing of administrative geographical names and research of spatial-temporal evolution.Therefore,based on the multi-source information on administrative geographical names,this thesis studies the construction method of knowledge graph for administrative geographical names.Through the extraction,fusion,and visualization expression of multi-source heterogeneous administrative geographical names,systematically sort out the information on the changes of administrative geographical names after 1949 and show the evolution process and laws of administrative geographical names in the past 70 years.The specific research content and achievements of this thesis are as follows:(1)Analysis of the characteristics of administrative geographical names based on multi-source data.Collecting administrative geographical names data from various sources such as the “People’s Republic of China Administrative Divisions Directory”,Local Chronicles of Various Provinces and Cities,Xingzheng Quhua Network,and Boya Geographical Name Network;analyzing the advantages and disadvantages of administrative geographical names data from various sources,picking authoritative,comprehensive and complementary data of administrative geographical names,forming evolution information on administrative geographical names,and analyzing the temporal characteristics,spatial characteristics and attributes of administrative geographical names and their evolution information.(2)Extracting method of information on evolution of administrative geographical names.The time information extraction method based on rule model,the method of extracting the toponyms information of administrative geographical names based on the integration of Bi-LSTM and CRF,the method of extracting the evolution relationship of the administrative geographical names based on the integration of Bi-LSTM and doublelevel attention,the semantic disambiguation method of administrative geographical names,semantic disambiguation method of administrative geographical names based on encyclopedia knowledge base and word vector and the attribute filling method of administrative geographical names based on attribute knowledge base are researched.The experimental results show that the time information extraction method based on the rule model has an accuracy rate of 99.12%,recall rate of 98.14%,F value of 98.63% in the evolution of the information of administrative geographical names;the accuracy rate,recall rate and F value of the method of extracting the information of administrative geographical names based on the integration of Bi-LSTM and CRF in the mixed corpus is 95.09%,93.17% and 94.12%;the method of extracting the evolution of administrative geographical names based on the integration of Bi-LSTM and double-level attention has an accuracy rate,recall rate and F value of 97.61%,93.76% and 95.65% in the evolution of the information of administrative geographical names.(3)The construction and application of the knowledge graph for administrative geographical names.The storing and visualizing of administrative geographical names based on the graph database Neo4 j is researched,and statistical analysis of administrative geographical names and analysis of spatial and temporal evolution pattern is conducted.Excavated the evolution laws and characteristics of administrative geographical names.Conduct querying and application for attribute information and evolution relationship of administrative geographical names based on the knowledge graph of administrative geographical names;and based on the 2015 Anhui province township division vector maps and the information on evolution relationship of administrative geographical names,carry out the spatial deduction for administrative geographical names above the county level in Anhui since 1949,construct the spatial-temporal database of administrative geographical names,the research carried out the application of querying the and geographical names,and develop a linkage platform for the visualization of spatial and temporal information of administrative geographical names based on knowledge graph,so as to realize the association and expression of the knowledge graph and temporal scope of administrative geographical names.