1、 2009 城镇居民收入与支出典型相关分析报告1、将数据导入 spss 形成项目文件：地 区工资性收入x1经营净收入x2财产性收入x3转移性收入x4食 品y1衣 着y2居 住y3医疗保健y4交通和通信y5教育文化娱乐服务y6其它商品和服务y7北 京 21105.611095.45586.727885.95936.111795.681290.221389.452767.852654.98833.32天 津 14389.1847.23305.318024.045404.531362.561505.71273.381968.371740.85634.05河 北 9830.57977.23193.71
2、4674.233250.771190.191142.83971.291151.15982.21361.83山 西 9741.38944.42252.394044.973071.93 11621319.45789.921095.771070.6281.61内蒙古 11267.41737.04363.813583.13772.631857.191246.21992.731557.031504.36641.96辽 宁 10420.61553.18239.815544.114680.851338.84 12931018.441493.171283.68609.09吉 林 9482.131307.311
3、45.734219.983637.321419.121394.941120.441305.451028.06465.42黑龙江 8356.661224.2988.944019.963397.411403.721026.77978.79922.77956.85395.41上 海 23172.361434.88473.47322.327344.831593.081913.221002.143498.653138.981136.06江 苏 13480.722139.83381.716492.694773.671297.951148.85808.371721.871968.03510.94浙 江 16
4、701.043294.481414.525709.265604.721614.661485.9984.623290.632295.32578.67安 徽 10362.391023.48272.874033.24051.41080.061219.83716.871013.381225.36337.36福 建 14211.492054.951172.764253.155336.361171.881394.91591.51993.771504.96598.13江 西 9789.791153.45239.833864.133881.561053.01935.44550.251145.161066.94
5、345.78山 东 13985.831379.02412.763559.33954.341548.751280.04885.161719.681332.97406.75河 南 9910.461202.69164.854130.053272.751270.741004.37875.521033.991048.14376.7湖 北 10331.511232.29296.633837.684160.511210.32999.49694.61953.691208.46307.75湖 南 9854.091744.37419.174060.494174.551146.251074.69784.661233
6、.821207.72408.14广 东 16898.882459.83736.554021.26225.221064.33 1814925.622979.882168.88627.01广 西 11193.641385.85493.363960.054129.55855.61021.11538.171598.681111.13343.33海 南 9678.651531.7424.463274.464507.81581.661000.32604.151548.76961.95296.28重 庆 118241018.76253.983893.574576.231503.491120.6982.731
7、189.031351.9377.02四 川 10132.431132.13305.383753.814391.731178.38973.02648.311416.491150.73422.38贵 州 9005.571135.05134.343518.423755.611012.14747.57535.43983.131146.35278.71云 南 9641.681092.291043.933902.384460.581102.14943.67708.781587.19798.69207.53西 藏 13326.4378.07218.31056.184581.61086.42689.76352
8、.311062.83465.84438.68陕 西 10775.37544.3152.33839.33988.571209.961018.23863.361071.481430.22440.35甘 肃 9182.24690.4159.492985.93359.31169.7801.21746.77894.351025.47334.95青 海 9341.26835.48 45.73927.813548.851043.4790.5701.37975.91889.32331.86宁 夏 9597.112036.14281.153636.363432.231260.581128.12921.86136
9、3.631075.88460.82新 疆 10232.91974.62115.772278.883386.331357.05856.78684.011198.65855.53436.72、实验步骤：在 file 中点击 new syntax，在语法窗口中输入程序：include D:学习软件 spsscanonical correlation.sps.cancorr set1=x1 x2 x3 x4/set2=y1 y2 y3 y4 y5 y6 y7.点击run菜单的all开始运行典型相关计算程序，运行结果如下：三、实验结果与分析：Run MATRIX procedure:图1、Correla
10、tions for Set-1x1 x2 x3 x4x1 1.0000 .3237 .4840 .6259x2 .3237 1.0000 .6878 .2635x3 .4840 .6878 1.0000 .2675x4 .6259 .2635 .2675 1.0000图2Correlations for Set-2y1 y2 y3 y4 y5 y6 y7y1 1.0000 .2525 .6899 .2970 .8827 .8281 .7826y2 .2525 1.0000 .4257 .7133 .3973 .5342 .6102y3 .6899 .4257 1.0000 .6058 .791
11、1 .7764 .7335y4 .2970 .7133 .6058 1.0000 .4331 .5692 .5571y5 .8827 .3973 .7911 .4331 1.0000 .8781 .7897y6 .8281 .5342 .7764 .5692 .8781 1.0000 .8494y7 .7826 .6102 .7335 .5571 .7897 .8494 1.0000图1和图2 给出的是两组变量内部各自的相关矩阵，图1知第一组变量间的有x1和x4相关系数为 0.6259，X2和x3之间为0.6878，说明变量间是存在相关性的；由图2可知，第二组变量之间也是存在一定的相关关系的。
12、图3.给出的是两组变量间各变量的相关矩阵，从此表可以看出每组变量间总有一个变量与另一组的变量有相关系数达0.6以上的，说明两组变量间确实存在相关性，这里需要做的就是提取出综合指标来代表这种相关性。Correlations Between Set-1 and Set-2y1 y2 y3 y4 y5 y6 y7x1 .8910 .4890 .6988 .4259 .8968 .8929 .8608x2 .3975 .1878 .5274 .1748 .6135 .4755 .3083x3 .5805 .1242 .4483 .0492 .6898 .4405 .2666x4 .6050 .4447
13、 .6349 .7219 .6219 .7800 .6548图4给出的是提取两个典型相关系数的大小，可见第一典型相关系数为0.982，第二典型相关系数为0.888，第三典型相关系数为0.582，第四个典型相关系数为0.511Canonical Correlations1 .9822 .8883 .5824 .511图5.检验的是各典型相关系数有无统计意义。第一对典型变量显著性检验的卡方统计量为134.264，p值为0，说明第一对典型相关变量显著相关。以此类推，第二对典型相关变量也显著相关；而第三对和第四对典型相关变量相关性不太显著，没有统计意义。因此选取1、2两对典型相关变量即可。Test t
14、hat remaining correlations are zero:Wilks Chi-SQ DF Sig.1 .004 134.264 28.000 .0002 .104 54.429 18.000 .0003 .489 17.159 10.000 .0714 .739 7.254 4.000 .123图6.显示各典型变量与变量组1中各变量间标化与未标化的系数图，由此可以写出典型变量的转化公式（标化的）：L1=0.764*x1 +0.241*x2+0.054*x3 + 0.143*x4 ； L2=-0.082*x1+0.073 *x2-0.869*x3+0.834*x4 ；以此类推可得L
15、3和L4的公式。Standardized Canonical Coefficients for Set-11 2 3 4x1 .764 .082 -1.177 .212x2 .241 -.073 .224 1.352x3 .054 .869 .602 -1.068x4 .143 -.834 .891 -.420Raw Canonical Coefficients for Set-11 2 3 4x1 .000 .000 .000 .000x2 .000 .000 .000 .002x3 .000 .003 .002 -.003x4 .000 -.001 .001 .000图7显示各典型变量与变
16、量组2中各变量间标化与未标化的系数图，同上可以写出典型变量的转换公式。M1=-0.180*y1-0.125*y2+0.011*y3+0.059*y4-0.532*y5-0.316*y6+0.01*y7；以此类推得M2、M3、M4。Standardized Canonical Coefficients for Set-21 2 3 4y1 .180 .284 -.213 -2.324y2 .125 .702 -.285 -.213y3 -.011 .250 .307 .706y4 -.059 -.869 .790 -.793y5 .532 1.339 .707 .727y6 .316 -1.05
17、4 .321 .721y7 -.010 -.811 -1.500 .763Raw Canonical Coefficients for Set-21 2 3 4y1 .000 .000 .000 -.002y2 .000 .003 -.001 -.001y3 .000 .001 .001 .002y4 .000 -.004 .004 -.004y5 .001 .002 .001 .001y6 .001 -.002 .001 .001y7 .000 -.004 -.008 .004图8显示第一组变量中个变量分别于自身、相关的典型变量的相关系数，可见他们主要和第一对典型变量的关系比较密切。Cano
18、nical Loadings for Set-11 2 3 4x1 .957 -.043 -.255 -.131x2 .563 .332 .492 .575x3 .627 .636 .425 -.148x4 .699 -.570 .375 -.217Cross Loadings for Set-11 2 3 4x1 .940 -.038 -.148 -.067x2 .553 .295 .286 .294x3 .616 .564 .247 -.076x4 .686 -.506 .218 -.111图9.显示第二组变量中个变量分别于自身、相关的典型变量的相关系数，可见他们主要和第一对典型变量的关系
19、比较密切。Canonical Loadings for Set-21 2 3 4y1 .910 .050 -.122 -.290y2 .498 -.266 -.107 .074y3 .790 -.136 .226 .226y4 .482 -.605 .363 -.058y5 .976 .124 .089 .041y6 .949 -.258 .027 .066y7 .855 -.300 -.344 .077Cross Loadings for Set-21 2 3 4y1 .894 .044 -.071 -.148y2 .489 -.237 -.062 .038y3 .776 -.121 .13
20、1 .115y4 .473 -.537 .211 -.029y5 .959 .110 .052 .021y6 .932 -.229 .016 .034y7 .840 -.266 -.200 .039下面输出的是冗余分析 Redundancy Analysis结果，它列出各典型相关系数所能解释的变量变异的比例，可以有了辅助判断需要保留多少个典型相关系数。由下图10 可知，第一组变量的差异可被自身的典型变量所解释的比例，其中第一典型变量占总变异的52.9%，而第二典型变量解释了21.0%。Redundancy Analysis:Proportion of Variance of Set-1 Exp
21、lained by Its Own Can. Var.Prop VarCV1-1 .529CV1-2 .210CV1-3 .157CV1-4 .104图11显示的是第一组变量的差异能被它们相对的典型所解释的比例，可见第一典型变量的解释读为51%，第二典型变量的解释度非常小。Proportion of Variance of Set-1 Explained by Opposite Can.Var.Prop VarCV2-1 .510CV2-2 .166CV2-3 .053CV2-4 .027图12、显示的是第二变量组的变异分别能被自身、相对的典型变量所揭示的比例，可见结论和上面一样，第二对变量的
22、贡献很小。因此综合分析冗余结果，我们只需要保留第一对典型变量即可。Proportion of Variance of Set-2 Explained by Its Own Can. Var.Prop VarCV2-1 .645CV2-2 .090CV2-3 .048CV2-4 .022Proportion of Variance of Set-2 Explained by Opposite Can. Var.Prop VarCV1-1 .622CV1-2 .071CV1-3 .016CV1-4 .006- END MATRIX -4、实验结论：1、L1=0.764*x1 +0.241*x2+0.054*x3 + 0.143*x4 2、M1=-0.180*y1-0.125*y2+0.011*y3+0.059*y4-0.532*y5-0.316*y6+0.01*y7
