第52卷第1期测绘学报Vol.52,No.12023年1月ActaGeodaeticaetCartographicaSinicaJanuary,2023引文格式:段晓旗,张彤,田有亮,等.居民出行异质性与城市活动结构[J].测绘学报,2023,52(1):155-166.DOI:10.11947/j.AGCS.2023.20210368.DUANXiaoqi,ZHANGTong,TIANYouliang,etal.Residents̓travelheterogeneityandurbanmobilitystructure[J].ActaGeodaeticaetCartographicaSinica,2023,52(1):155-166.DOI:10.11947/j.AGCS.2023.20210368.居民出行异质性与城市活动结构段晓旗1,2,张彤2,田有亮1,刘沛林3,万桥1,秦永彬11.贵州大学计算机科学与技术学院,贵阳550025;2.武汉大学测绘遥感信息工程国家重点实验室,武汉430079;3.长沙学院乡村振兴研究院,长沙410022Residents̓travelheterogeneityandurbanmobilitystructureDUANXiaoqi1,2,ZHANGTong2,TIANYouliang1,LIUPeilin3,WANQiao1,QINYongbin11.ComputerScienceandTechnologyInstitute,GuizhouUniversity,Guiyang550025,China;2.StateKeyLaboratoryofInformationEngineeringinSurveying,MappingandRemoteSensing,WuhanUniversity,Wuhan430079,China;3.SchoolofEconomicsandManagementandRuralVitalizationInstitution,ChangshaUniversity,Changsha410022,ChinaAbstract:Theurbanmobilitystructurecouldproducedynamicchangesintheprocessofresidents̓travel.Discoveringurbanmobilitystructurehasbecomeahotissueinurbangeographyandtrafficgeography,whichisofgreatsignificanceforurbanplanningandmanagement,lineallocation,trafficestimation,etc.Traditionalurbanstructurediscoverymethodsusuallyconsiderasinglefactor,andlackofresearchinconsideringtheheterogeneityofresidents̓travel.Basedontheauto-encodermodel,thispaperproposesarepresentationallearningmethodconsideringtheheterogeneityofresidents̓traveltodetecttheurbanmobilitystructure.OurmethodrealizesthefusionofstaticattributeinformationanddynamictravelinformationandintroducesthemultivariateGaussiandistributionmodeltocompletethewholeprocessfromnodeexpressiontocommunityexpressionconsideringtheresidents̓travelheterogeneitybetweendifferentcommunities,finallycanachievemoreaccuraterepresentationalresults.Theexperimentalresultsshowthattheproposedmethodcanmoreaccuratelydiscovertheurbanmobilityst...
