Spatio-temporal Co-location Pattern Mining And Applications In Urban Congestion Identification

In recent years,with the popularization of remote sensing technology and mobile devices,data such as remote sensing data and urban traffic data show explosive growth.These huge amounts of data enable people to discover hidden knowledge from data and apply them to related fields,such as life service,location recommendation,traffic trip and so on.In the field of transportation,due to the rapid development and expansion of cities,the number of cars has increased dramatically in cities.Urban traffic congestion has become an urgent problem in today's social development.Therefore,effective excavation and analysis of urban traffic data,providing directly available knowledge,is of great significance for reducing traffic congestion probability and ensuring people travel to avoid peak areas and improve travel efficiency.Traditional traffic congestion algorithm only considers traffic congestion on a section.However,in urban traffic,traffic congestion often has transitivity.Congestion of a road often causes congestion around many roads.Considering the correlation of spatial data distribution,this paper introduces the co-location pattern mining method into the research of traffic congestion pattern mining.However,the traditional co-location pattern usually focuses on the neighborhood relationship between spatial features.Traffic congestion analysis also takes into account many factors such as time information,vehicle speed and so on.Firstly,this thesis introduces the related knowledge and basic concepts of traditional co-location pattern mining;Secondly,based on the traditional co-location pattern mining,considering time,vehicle speed and other property,and consider the transfer of the characteristics of traffic congestion,came up with a new concept of spatio-temporal congestion co-location pattern.In view of the transitivity characteristics of traffic congestion,a new concept of space-time co-location congestion model is proposed,and a series of relevant definitions are given.;Third,an effective spatio-temporal congestion co-location pattern mining algorithm is designed and a corresponding pruning strategy is given.In addition,this paper designs the reliability index,and gives the pattern results for further analysis,mining directly available knowledge;Fourth,through a large number of experiments on the traffic data set in Guiyang City,the correctness and versatility of the algorithm proposed in this paper are verified.Reliability index analysis of the results obtained from the experiments is made to obtain directly usable knowledge,thus helping traffic grooming and planning.Finally,we summarize the work of this article,and at the same time,we also look forward to the future work.