Design And Implementation Of Multi Spatial Entity Linkage System For Geographic Information Retrieval

With the increasingly rich information of geographic entities,in addition to navigation,There are scenes with more restrictions on target features such as factory site selection and intelligence personnel assembly on demand.At present,main map service providers can only provide fuzzy retrieval based on target name.Searching with multiple constraints can only be performed by manual sub-conditional screening one by one,which is time-consuming and inefficient,and sometimes the attribute does not exist in the database which means the search cannot be completed at all.So,the current core problem of the multi-conditional geographic retrieval scenario lies in the lack of data volume and data attributes,while the current geographic data sources are diverse and can merge each other to enhance data quality.This article studies the multi-source geographic data fusion task,designing and implementing a multi-source data fusion system,which collects and fuses geographic data,and finally implements complex conditional geographic query functions.First,this paper proposes a spatial indexing entity linkage(SIEL)geographic entity fusion algorithm based on the Dedupe entity linkage algorithm for geographic data.The algorithm follows the basic idea of Dedupe,which regards the fusion problem as a classification problem of whether the candidate entity pair links the same geographic entity.On this basis,the spatial feature is added to establish an index to reduce the size of the partition,and the similarity measure is improved based on the geographic attribute characteristics.A decision tree classifier with better classification performance instead.Tested on Gaode and Google Maps datasets,the results prove that SIEL has a 11.4%improvement in the F measure compared to the Dedupe algorithm.Finally,based on the geographic data fusion algorithm,this paper designs and implements a multi-source data fusion system for geographic information retrieval.The system can regularly collect online map data,and then clean it and fuse it into a unified data model for persistence.The query is decomposed and executed according to the designed composite query model,and multiple conditional results are combined to complete the composite conditional geographic query function.The test proves that the system can effectively perform data fusion and complete the complex conditional geographic query function,which proves the feasibility and effectiveness of the system.