用聚类分析分析各个城市的差距

《用聚类分析分析各个城市的差距》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

1、用聚类分析分析各个地区的差距及将比较相近的地区划归为一类使用数据如下表所示，利用聚类分析将下面28个地区划分为4类，并比较四类地区之间的差异。地区序号食品衣着燃料住房生活用品文化牛•活类别天津1135.2036.4010.4744.1636.403.941辽宁2145.6832.8317.7927.2939.093.471吉林3159.3733.3818.3711.8125.295.221江苏4144.9829.1211.6742.6027.305.741浙江5169.9232.7512.7247.1234.355.001山东6115.8430.761

2、2.2033.6133.773.851黑龙江7116.2229.5713.2413.7621.756.049安徽8153.1123.0915.6223.5418.186.392福建9144.9221.0616.9619.5221.756.732江西10140.5421.5917.6419.1915.974.942湖北11140.6428.2612.3518.5321.956.232湖南12164.0224.7413.6322.201&066.042广西13139.0818.4714.6813.4120.663.852四川14137.8020.7411.

3、0717.7416.494.392贵州15121.6721.5312.5814.4912.184.572新疆16123.243&0013.724.6417.775.752河北1795.2122.839.3022.4422.812.803山西18104.7825.116.469.8918.173.253内蒙19128.4127.638.9412.5823.993.273河南20101.1823.268.4620.2020.504.303云南21124.2719.818.8914.2215.533.033陕西22106.0220.5610.9410.111

4、8.003.293甘肃2395.6516.825.7()6.0312.364.493青海24107.1216.458.985.408.785.933宁夏25113.7424」16.469.6122.922.533北京26190.3343.779.7360.5449.109.044上海27221.1138.6412.53115.6550.825.894广州28182.5520.521&3242.4036.9711.684使用聚类分析所得结果如2表一InitialClusterCentersCluster1234食品95.65221.11164.02190

5、.33衣着16.8238.6424.7443.77燃料5.7012.5313.639.73住房6.03115.6522.2060.54生活用品12.3650.8218.0649.10文化生活4.495.896.049.04表一是K・均值聚类分析的初始类中心数据，总共是四类。表二IterationHistory*IterationChangeinClusterCenters1234120.892.00019.11520.6122.000.000.000.000a.Convergenceachievedduetonoorsmallchangeinclust

6、ercenters.Themaximumabsolutecoordinatechangeforanycenteris・000.Thecurrentiterationis2.Theminimumdistaneebetweeninitialcentersis59.262.表二是迭代记录，从表中可以看出，第一次迭代后，四个类的中心点分别变化了20.895,0,19.115和20612,第一个中心点的变化最大，第二次迭代后，4个中心点变化均小于指定的收敛准则（0.01）,达到聚类结果的要求，聚类分析结束。表二ClusterMembershipCaseNumbe

7、r地区ClusterDistance1天津342.4212辽宁329.8983吉林316.7224江苏329.7855浙江430.9996山东143.1707黑龙江128.1588安徽311.2989福建320.26610江西324.33011湖北324.27512湖南3・00013广西327.40914四川327.11315贵州12&60416新疆136.15117河北120.75118山西114.24319内蒙137.15120河南162121云南130.29022陕西114.10223甘肃1.00024青海112.56025宁夏122.56426

8、北京4・00027上海2・00028广州434.030表三显示聚类的类成员列表，给出了最终K■