Mining Local Tight Spatial Sub-prevalent Co-location Pattern

Today,society has entered the information age,and various fields will generate a large amount of spatial data when facing the transformation of informationization.Therefore,spatial data mining has become a research direction that cannot be ignored.In spatial data mining,rapid and efficient spatial co-location pattern mining has received great attention from researchers in recent years,and has yielded rich results.The mining goal of co-location pattern is to find the spatial things that appear together frequently,so the identified co-location pattern only reflects the co-occurrence relationship between spatial things;The traditional co-location pattern based on cluster instance model considers whether the instances can form clusters,but ignores some important neighbor relationships that are not clusters.In response to the above problems,this paper chooses the star instance model as the basic model,studies two co-location patterns that consider both frequency and spatial distribution,and designs and implements a prototype system based on these two modes.The main work is as follows:(1)A local tight sub-frequent co-location pattern mining is proposed.Analyze real-world spatial data and find that the same frequent co-location patterns will have different distribution characteristics.For example,some frequent patterns in space are closely distributed in a certain area,which is more practical and easier to explain.Therefore,this paper proposes local tight sub-frequent co-location pattern mining.Firstly,this thesis designs the relevant definitions of local tight patterns,then study the algorithm of mining the pattern,and the pruning conditions of the optimization algorithm.,Finally,through a large number of experiments,the efficiency of local tihgt pattern mining algorithm is analyzed,and the practicability of local tight pattern is verified.(2)A local tight spatiotemporal sub-prevalent co-location patterns mining is proposed.In order to analyze the temporal nature of spatial data,time is integrated on the basis of the above research,and a more meaningful co-location pattern is found.Firstly,this thesis solves the problem of introducing time characteristics into local tight co-location pattern mining.Then,a local tight spatio-temporal sub-frequent co-location pattern mining method is designed.Finally,experiments are carried out on the test data set to evaluate the mining efficiency and mining results of the proposed algorithm.(3)Design and implement a prototype system for local tight co-location pattern mining and analysis.The system is developed through the Py Qt5 module in Python,and displays three functional modules: local tight spatial co-location pattern mining,local tight spatiotemporal co-location pattern mining,and mining result visualization.