Mining Constraint-based Spatial Co-location Patterns And Applying In Urban Planning

With the continuous development of the world's economy,the urban scale is expanding,and the problem of unreasonable urban construction is becoming more and more serious.Therefore,it is particularly important to find effective methods to improve the unreasonable distribution of urban facilities.In recent years,spatial data has gradually become the basis and subject of various information systems.Spatial data mining technology has been widely applied while developing at a high speed.Spatial co-location pattern mining,as an important research direction of spatial association rule mining,has been vigorously promoted in the application of urban planning.However,the general research results are based on the idealized global spatial data,but ignore the necessary constraints of urban planning in real life.In order to make the mining results more fit with the reality of life,this paper,by using the constrained spatial co-location pattern mining algorithm,studies the reasonable advanced urban layout planning and obtains the distribution law of the featured objects,so as to help the relevant government agencies to carry out the development of new urban areas or the overall planning of new cities.First,this paper extracts the coordinates of the elements in Singapore,a mature city with reasonable layout to be analyzed,and then compiles relevant programs to prune and process the spatial elements to be mined using the join-base algorithm in co-location pattern mining,and obtains the processing results.Firstly,this paper summarizes the current research status of data mining and urban planning,introduces the basic concepts of spatial association rules and their mining algorithms,then introduces the spatial co-location mining mode,and expounds its concepts,mining methods and current research results.Secondly,this paper uses mining method with constraints,to dig the mature urban Singapore:some different characteristics between objects in co-the location pattern mining,get all meet the conditions of model,on the basis of the join constraints,pruning of the data processing,the resulting urban planning mining results conform to reality.At the end of the paper,the mining results are analyzed,and compared with the mining results without constraints,and the conclusion is drawn.The author takes residential area as the core characteristic to carry on the restraint,preserves its correlation frequent pattern,on this foundation deletes processing to some specific pattern,then carries on the frequent pattern mining.In data processing,this paper innovates in calculating spatial proximity R,using the formula of calculating the precise distance between two points on the earth by latitude and longitude,which makes the distance data more precise.On algorithm,some optimizations are made based on the classical Join-based algorithm,proposed E-C Join-based algorithm,that is,in its operation to generate candidate model of step,to join the constraint conditions,delete does not meet the constraint conditions of model,matching the constraint conditions satisfy min_prev and min_conf co-rules of the location of all the space.In this way,the mining results can not only satisfy the constraint conditions,but also reduce the following unnecessary calculations,thus improving the efficiency of the algorithm and making the mining results more valuable for reference in real life urban planning.