Font Size: a A A

Delimiting Vague Places Using Point Of Interest Datasets

Posted on:2019-07-14Degree:MasterType:Thesis
Country:ChinaCandidate:F YanFull Text:PDF
GTID:2370330545999722Subject:Cartography and Geographic Information Engineering
Abstract/Summary:PDF Full Text Request
A place name is an exclusive name given by a person to a natural or humanistic geographical entity in a particular spatial position.Places names are widely and frequently used in daily life when people describe the locations of places,give some navigational instructions,or notify the places of activities.Most of places people use are vague and have no limit geospatial boundaries that are officially accepted,and many gazetteers lack of definite representations for boundaries of vague places.Therefore,delimiting vague places can not only enrich gazetteers but also offer more efficient and more accurate geographical information services.To determine a crisp boundary of a vague place,an automated method is proposed in this research.The appropriate value of Eps in DBSCAN is calculated by 6-dist graph,and then DBSCAN is implemented for POI datasets of vague places,the convex hull of the cluster that contains the maximum number of points is decided to be the approximate boundary of a vague place.The proposed method avoids manual intervention,which reduces time and effort costs in large-scale application.A method based on rough set theory for evaluating regions of vague place is proposed in this research.KDE,grid calculation method,ASCDT and distance analysis method are performed to generate four regions for a vague place,and then the union and intersection of the four regions are the upper and lower approximation for the vague place.Two indicators are defined to reveal that to what extent a region overlaps with the upper and lower approximation of a vague place.In the experiments of modelling 24 vague places of Wuhan city,little difference is observed in the evaluations for results generated by automated method and by manual intervention,which shows that the automated method can replace manual intervention.To understand how the regions of vague places change,a questionnaire is designed to acquire people’s cognition about 5 vague places.It is found that people’s familiarity with a place influence their cognition about the region of the place.The differences between the regions delimited by the automated method and people’ cognitive regions reveal the changes of vague places,and show that the expansion of roads may enlarge vague places,and name borrowing causes gaps between generated results and cognitive regions and may mislead people’s cognition about vague places.
Keywords/Search Tags:vague places, spatial cognition, clustering, point of interest
PDF Full Text Request
Related items