Font Size: a A A

Delimitating Vague Place Name Regions Based On Outlier Removal

Posted on:2018-07-02Degree:MasterType:Thesis
Country:ChinaCandidate:X Y HuangFull Text:PDF
GTID:2370330512482863Subject:Cartography and Geographic Information Engineering
Abstract/Summary:PDF Full Text Request
Place names used in daily life can indicate the locations of regions,and is a way of spatial information communication.Place names also play an important role in describing and retrieving geographic information in geographic information systems.However,there are only a part of place names that have definite boundary,such as place names of administration district.Many vague place names used everyday has no clear boundary,due to the vagueness of hunman cognition generated when human beings perceive and conceptualize the world.Digital gazetteer stores the information of place names by the triple<name,type,space scope>,and provides basic data service for GIS.But there is often a lack of accurate boundary information of vague place names in gazetteer.Therefore,research on the scope of vague place names can enrich gazetteer and extend its applications,and provide more efficient GIS service to the public.Now the applications of Big Data supplies new methods for delimitating Vague Places.According to the shorcomings of state-of-the-art,a new method of of delimitating vague place name is proposed in this paper.This method uses Sina Weibo checkin data to define the boundaries of vague place name regions.The method combines outlier detection algorithm and Gaussian mixture model,which means that the probability density distribution approximation of checkin POIs outlier degrees is established using Gaussian mixture models.Consequently,the outliers in checkin POIs are detected,and the boundary is defined based on the rest POIs.In addition,we analyse the performance of both KNN algorithm and LOF algorithm.In a word,the proposed method can eliminate POIs that may not in the region of the place names,and is less insensitive to the parameter.In this paper,the boundaries of 6 districts and 6 vague place names regions in Wuhan are defined based on the outlier removal method.The 6 districts are mainly used for quantitative evaluation and demonstrate the feasibility of the proposed method,and the 6 vague place names regions verdict the validity of this method.
Keywords/Search Tags:vague place names, outlier detection, spatial cognition
PDF Full Text Request
Related items