Font Size: a A A

Recommendation Of Complaint Address Based On Spatial Scene Similarity

Posted on:2018-02-22Degree:MasterType:Thesis
Country:ChinaCandidate:H X WanFull Text:PDF
GTID:2359330515997744Subject:Cartography and Geographic Information Engineering
Abstract/Summary:PDF Full Text Request
With the development of geographic information technology,massive spatial data blowout,the traditional data retrieval model has been unable to meet people's demand for spatial data query.For example,to locate an address,people often enter the name of address to query,but if the name is entered in error or does not match the name in spatial database,it is usually not possible to obtain the desired result.Gestalt Psychology pointed out that people's cognition of space is from the whole to the part.Based on this cognitive process and people's description habit of space,this paper proposes an address matching model based on spatial scene similarity.In the case that user does not know the exact name of the address to be queried,by describing the spatial relationship of the query address,the address that corresponds the spatial relationship can be retrieved and be further selected.Spatial data retrieval based on spatial similarity is more close to the way of thinking about spatial relations,which is of great significance to the future intelligent retrieval of spatial data.In order to achieve the rapid recommendation of 12315 complaint address,this paper studies the characteristics and calculation methods of spatial scene similarity.The spatial scene characteristics include two aspects:the characteristics of spatial target and spatial relations.In the study of spatial target characteristics,this paper proposes a method based on the combination of semantic similarity between type ontology and literal similarity based on editing distance,which retrieves the geographical information input by user in order to solve the problem that the exact match can not get the query result.The similarity calculation of spatial relations mainly includes spatial azimuth,topology and distance relation.This paper summarizes the existing similarity calculation methods of three kinds of spatial relations,and puts forward the spatial relationship calculation model adopted in this paper.Finally,based on the similarity between spatial target and spatial relation,a spatial scene calculation model considering local importance is proposed,and the model is validated in the prototype system.The main researches and results are as follows:(1)A local name matching model based on semantic similarity between type ontology and literal similarity based on edit distance is proposed.In order to solve the problem that the address of 12315 complaint hotline can not match exactly in 12315 complaint hotline,this paper constructs the ontology of the type,and combines the similarity degree and the semantic similarity of the type to put forward the name and address matching model.(2)Propose spatial similarity calculation method.In this paper,the spatial relationship includes spatial azimuth,topology and distance relation.Based on the existing spatial relationship similarity calculation method,this paper proposes a spatial similarity calculation model for the application scenario and data characteristics of the 12315 complaint hotline.(3)Based on the spatial target characteristics and spatial relation,a spatial scene similarity calculation model considering local importance is proposed.Considering that the degree of the influence of different spatial objects on spatial similarity may be different,this paper improves the similarity calculation model of spatial scene,and allows the user to define the influence level of space object on spatial similarity.(4)Develop prototype system.Based on the analysis of the requirement of address rapid positioning in 1235 complaint hotline,this paper proposes a solution based on spatial scene similarity,and develops a prototype system to verify the feasibility of the proposed similarity calculation model.
Keywords/Search Tags:type ontology, spatial relation, target characteristics, spatial scene similarity, address recommendation
PDF Full Text Request
Related items