空间权重矩阵对空间自相关的影响分析_以湖南省城乡收入差距为例

20102Feb.2010()JOURNALOFSOUTHCHINANORMALUNIVERSITY(NATURALSCIENCEEDITION)20101No.1,2010:2009-09-23:(40871086):(1981),,,2007,Email:[e-mail];(1969),,,,:,Email:[e-mail].3:1000-5463(2010)01-0110-06,3,(,510631):,.Arcgis913,.,,MoranI,.:;;MoranI:K903:A,[1].,.,,.,[2]ArcInfo;[3]01;[4]ArcobjectRooks01.,,,;[5],,MoranI,,MoranI.ANSELIN[6]GeoDaEsriArcMapSpatialSta2tisticsTools.2,.Arc2Map,.SpatialStatisticsTools,,.11.1,2,,[7].1.1.1MoranI(MoranI),[8-9]:I=n6ni=16nj=1wij6ni=16nj=1wij[xi-€x][xj-€x]6ni=1[xi-€x]2(Pji),(1)n,wij,xi,€x.MoranI(-1,1),,I0,I,,.I,,Z:ZI=I-E[I]Var[I],(2)E[I],Var[I].1.1.2MoranI,.:Ii=xi-€x6nj=1,jiwijn-1-€x26nj=1,jiwij[xi-€x].(3)Ii[8-9].1.2ArcGIS9.3(1)(3),wij.ArcGIS9.3GenerateSpatialWeightsMatrix.2(1).1ArcGIS9.3Tab.1ThemethodsappliedtogeneratespatialweightmatrixinArcGIS9.3BinarystrategiesFIXED_DISTANCEK_NEAREST_NEIGHBORSCONTIGUITY_EDGES_ONLYCONTIGUITY_EDGES_CORNERSDELAUNAY_TRIANGULATIONWeightedstrategiesINVERSE_DISTANCEZONE_OF_DIFFERENCEBinarystrategies,01;Weightedstrategies,01.1.2.1WeightedstrategiesWeightedstrategies,[10]:wij=1dk(i,jd);(4)wij=0().(5)INVERSE_DISTANCE(ID):,k12.,d(Thresholddistance)N(Num2berofneighbor).d3:Null0.,,0,(4);Null,d1(Neighbor);0,d.N.ZONE_OF_DIFFERENCE(ZD):,d0,d,(4).d,1.,,,,[10].1.2.2BinarystrategiesBinarystrategies,wij=1(i,j)wij=0().FIXED_DISTANCE(FD):dN.d0.d,1,0;d,1.K_NEAREST_NEIGHBORS(KNN):K,KK,K1,0.CONTIGUITY_EDGES_ONLY(CE):,[11],1,20.,N.N,,.CONTIGUITY_EDGES_CORNERS(CEC):CE,,1,0.DELAUNAY_TRIANGULATION(DT):,2,1,0.,,,.22.12007881111:,,..2008.2007(1).1,2007,,3.,8833,,2902.7106883313,6107710625,5244610725,4224524419,290242246.1Fig.1Thespatialdistributionofincomegap2.2,.2,:ID:d=0m,87,N;d,1,1d1,d1,7.N.N7,d1d2,N,N;N7,d1,N,7;1,CalculateDistanceBandfromNeighborCount.10,:74413m,153344m,101432m;d=153344m,,10,32.N11,;10N32,N,32;N32,N.FD,d=74413m,,10.N10,N;N11,10,N.KNN,N=10,10.CE/CECDT,,2.2.3,.2.3.18(3).,,MoranI.,ID,d=0m,MoranI.FD,10,,MoranI1.MoranI.,2,Null,N,MoranI.,(3.95),MoranI,,,2MoranI;,IDMoranIFD(2).211()20102Tab.2Theneighborsbydifferentpreferencesd/mN0()////ID(k=1,)0N88878787NullN8N7NullN7NNN153344N11103221.1415334410N32N32153344N32NNNFD74413N11N1074413N10NNNKNNNull10101010CENullNull194.68CECNullNull194.68DTNullNull395.643(=0.01)Tab.3TheeffectofspatialweightmatrixonGeneralMoranI(=0.01)dNMoranIZscoreID0Null0.04894.811Null100.21744.902153344100.15004.778FDNull100.20935.148153344100.11164.137KNNNull100.20935.148CE/CECNullNull0.27144.002DTNullNull0.23633.9962Fig.2TheaverageneighborsandGlobalMoranIKNNFD,N7,.DTCE/CEC,,,DTCE/CEC,CE/CEC(2),MoranICE/CECMoranI.2.3.24,43.4(=0.01)Tab.4TheeffectofspatialweightmatrixonLocalMoranI(=0.01)dNHHHLLHLL3(a)CE/CECNullNull31183(b)IDNull80621103(c)FDNull8016003(d)FDNull8701404HH(),HL(),LH(),LL(),(3).3(a)3(d)4.3(a)3(b)1LH,,,;1HL,;LL,(1);HH,3HH,(1),6.3(c)3(a)3(b),LL3(c).3(d),HHLL.,2.3.1,FDID,Null,N,3111:.4.3Fig.3TheresultsofLocalMoranI4Fig.4Thenumberofclustersandaverageneighbors4(a),,2HHLL,IDLLFD;,IDHHFD.4(b),,2HLLH;,FDHLLLID.CEC/CEDT,KNNFD,.411()20103CONTIGUITY_EDGES_CORNERS,.,,MoranI,4.,,,.MoranI,,GearysCG,.:[1],,.[J].,2008,20(4):317.HEZonggui,HANShimin,CUIDao.Discussiononstatis2ticofspatialautocorrelationanalysis[J].ChinaSchistoseControl,2008,20(4):317.[2],.[J].,2002(2):38-44.LIUXuhua,WANGJinfeng.Analysisandapplicationonthespecificationmethodsofthespatialweightmatrix[J].Geo-InformationScience,2002(2):38-44.[3],.ArcGIS[J].,2005,28(6):78-79.WEIXiaofeng,WUJianping.TheDevelopmentandappli2cationofspatialautocorrelationbasedonArcGIS[J].Geomatics&spatialinformationtechnology,2005,28(6):78-79.[4],,.ArcObjects[J].,2007,32(6):130-132.PANHaiyan,CHENGPenggen,XIAOGenru.Theconsti2tutionandrealizationofspatialweightmatrixbasedonAr2cObjects[J].ScienceofSurveyingandMapping,2007,32(6):130-132.[5].MoranI[D].:,2007:2-19.[6]ANSELINL.ExploringSpatialDatawithGeoDaTM[EB/LO].(2004-05-05)[2009-07-20].[7],,.[J].,2005,24(6):956.MENGBin,ZHANGJingqiu,WANGJinfeng.Applicationofspatialanalysistotheresearchofrealestate:takingBeijingasacase[J].Geographyresearch,2005,24(6):956.[8],.[M].:,2007:110-119.[9].[M].:,2006:120-126.[10]ESRI.ArcGISDesktopHelp[EB/LO].(2009-04-24)[2009-07-20].=welcome.[11]LEEJ,WONGWS.StatisticalanalysiswithArcViewGIS[M].NewYork:JohnWiley&Sons,Incorporated,2000:132-133.ANALYZINGTHEEFFECTOFSPATIALWEIGHTEDMATRIXONSPATIALAUTOCORRELATIONTAKINGHUNANpSINCOMEGAPBETWEENURBANANDRURALAREASASACASEWANGHongliang,HUWeiping3,WUChi(SchoolofGeography,SouthChinaNormalUniversity,Guangzhou510631,China)Abstract:Spatialweightmatrixquantificationallyexpressesthespatialrelationshipsbetweenthegeographicfea2tures,whichisanimportanttoolforspatialstatisticalanalysis.Thedifferentmethodsingeneratingspatialweightmatrix,whichisusedforspatialautocorrelationanalysisinArcGIS9.3,areanalyzedbyHunanpscountyincomegapbetweenurbanandruralasacaseforempiricalanalysis.Theanalysisindicatesthatwiththeexpansionofaverageinfluencedscopeofeachfeature,theglobalMoranIsh