Research On Place Name Information Retrieval Method Of Geographic Entity

With the advancement of digital city construction and the popularization of online map application,searching and retrieval for massive place names have become the core function of all kinds of city-level GIS application platform and online map platform.Place name information retrieval has become a hot issue in the related study area.A lot of research results about this issue has been found,however,these researches,which usually focused on the efficiency of retrieval and the matching quality of text information,did not take into account of the contradiction of the actual application requirements and place name retrieval status.Currently,many online map platforms generally take POI or point of place name as retrievable objects,and through retrieving text keywords to get results.However,this way ignores the spatial information of place name,and is difficult to achieve the complex retrieval request.The key to solve these problems is communal development of retrieve data and retrieval methods.In recent years,the online map,represented of Map World,began to focus on geographic entity data,but not proposed to apply these data into place name information retrieval.Geographic entity and place name has close relationship naturally.This relationship provides the conditions and opportunities for the expansion of retrieval data and development of retrieval methods.Therefore,this paper to study the forms and methods of place names information retrieval based on the attribution and spatial information,through combine geographic entity and place name.In this paper,the main research contents and results are as follows:(1)Analyze the data models,data structures and relationships of geographic entity and place name,and on this basis analyzes the impact of online map place names retrieval application of ignoring geographical entity.this paper proposes a new place name information retrieval mode,which based on geographic entity,and achieved the structural expression of it.This retrieval mode takes geographic entities'attribution and spatial relationship to be the retrieval feature items,which make a foundation for the study of retrieval model and method.(2)Relying on the place name information retrieval mode which take acount of geographic entity,and referencing the vector space model and geographic information retrieval model,this paper proposes a place name information retrieval model based.on geographic entity similarity(GS_PNIR).Through study the attributes feature type of the objects to retrieve and classification of spatial relationships,forming the corresponding similarity calculation method of attribute similarity and spatial relationship similarity.This model is suitable for retrieving place name information based on text keywords and spatial relationship.(3)Using GS_PNIR model and Lucene,this paper constructs a place name information retrieval framework.The paper realizes the construction of spatial index with the help of Lucene Spatial and make it possible to organize and manage text index together with spatial index.In addition,the paper improves Lucene's similarity calculation model by GS_PNIR model and on this basis designs the algorithm of place information retrieval and sorting,which is feasible verified by experiment.Results in this experiment show that,this retrieval method can apply to all kinds of retrieval situations which based on attribution and spatial relationship,and improve the accuracy of search results as well as the quality of sorted result.