實驗六--各國家和地區(qū)男子徑賽記錄數(shù)據(jù)的聚類分析

1、實驗六 各國家和地區(qū)男子徑賽記錄數(shù)據(jù)的聚類分析一、實驗目的1.掌握如何使用SAS軟件來進行聚類分析；2.看懂和理解SAS輸出的結(jié)果，并學會以此來作出分析；3.對同一組數(shù)據(jù)使用五種系統(tǒng)聚類方法，學會對各種聚類效果的比較，獲取重要經(jīng)驗；二、實驗內(nèi)容實驗1 下表中列出各國家和地區(qū)男子徑賽記錄的數(shù)據(jù)，試分別用類平均法、離差平方和法和k均值法進行聚類分析，聚類前先對各變量做標準化變換。國家和地區(qū)Alphabet100米（秒）200米（秒）400米（秒）800米（秒）1500米（秒）5000米（秒）10000米（秒）馬拉松阿根廷a10.3920.8146.841.813.714.0429.36137.72

2、澳大利亞a10.3120.0644.841.743.5713.2827.66128.3奧地利a10.4420.8146.821.793.613.2627.72135.9比利時b10.3420.6845.041.733.613.2227.45129.95百慕大b10.2820.5845.911.83.7514.6830.55146.62巴西b10.2220.4345.211.733.6613.6228.62133.13緬甸b10.6421.5248.31.83.8514.4530.28139.95加拿大c10.1720.2245.681.763.6313.5528.09130.15智利c10.3

3、420.846.21.793.7113.6129.3134.03中國c10.5121.0447.31.813.7313.929.13133.53哥倫比亞c10.4321.0546.11.823.7413.4927.88131.35庫克群島c12.1823.252.942.024.2416.735.38164.7哥斯達黎c10.9421.948.661.873.8414.0328.81136.58捷克斯洛c10.3520.6545.641.763.5813.4228.19134.32丹麥d10.5620.5245.891.783.6113.528.11130.78多米尼加d10.1420.654

4、6.81.823.8214.9131.45154.12芬蘭f10.4320.6945.491.743.6113.2727.52130.87法國f10.1120.3845.281.733.5713.3427.97132.3德意志民g10.1220.3344.871.733.5613.1727.42129.92德意志聯(lián)g10.1620.3744.51.733.5313.2127.61132.23大不列顛g10.1120.2144.931.73.5113.0127.51129.13希臘g10.2220.7146.561.783.6414.5928.45134.6危地馬拉g10.9821.8248.4

5、1.893.814.1630.11139.33匈牙利h10.2620.6246.021.773.6213.4928.44132.58印度ii10.621.4245.731.763.7313.7728.81131.98印度尼西i10.5921.4947.81.843.9214.7330.79148.83以色列i10.6120.9646.31.793.5613.3227.81132.35愛爾蘭i10.712147.81.773.7213.6628.93137.55意大利i10.0119.7245.261.733.613.2327.52131.08日本j10.3420.8145.861.793.64

6、13.4127.72128.63肯尼亞k10.4620.6644.921.733.5513.127.38129.75韓國k10.3420.8946.91.793.7713.9629.23136.25朝鮮人民k10.9121.9447.31.853.7714.1329.67130.87盧森堡l10.3520.7747.41.823.6713.6429.08141.27馬來西亞m10.420.9246.31.823.814.6431.01154.1毛里求斯m11.1922.4547.71.883.8315.0631.77152.23墨西哥m10.4221.346.11.83.6513.4627.9

7、5129.2荷蘭n10.5220.9545.11.743.6213.3627.61129.02新西蘭n10.5120.8846.11.743.5413.2127.7128.98挪威n10.5521.1646.711.763.6213.3427.69131.48巴布亞新p10.9621.7847.91.94.0114.7231.36148.22菲律賓p10.7821.6446.241.813.8314.7430.64145.27波蘭p10.1620.2445.361.763.613.2927.89131.58葡萄牙p10.5321.1746.71.793.6213.1327.38128.65羅馬

8、尼亞r10.4120.9845.871.763.6413.2527.67132.5新加坡s10.3821.2847.41.883.8915.1131.32157.77西班牙s10.4220.7745.981.763.5513.3127.73131.57瑞士s10.2520.6145.631.773.6113.2927.94130.63瑞典s10.3720.4645.781.783.5513.2227.91131.2臺北t10.5921.2946.81.793.7714.0730.07139.27泰國t10.3921.0947.911.833.8415.2332.56149.9土耳其t10.71

9、21.4347.61.793.6713.5628.58131.5美國u9.9319.7543.861.733.5313.227.43128.22蘇聯(lián)u10.072044.61.753.5913.227.53130.55西薩摩亞w10.8221.86492.024.2416.2834.71161.83三、實驗要求1.完成三種系統(tǒng)聚類方法，比較其聚類效果；2.能對試驗結(jié)果進行合理分析.四、實驗內(nèi)容4.1完成三種系統(tǒng)聚類方法，比較其聚類效果4.1.1導入數(shù)據(jù)集exec654.1.2類平均法聚類proc cluster data=work.exec65 method=ave nosquare std;

10、 id nation;proc tree horizontal; id nation;run; 答：用類平均法聚類的話認為較合適的是聚成4類，但分成兩類以上的話，庫克群島和西薩摩亞都會單聚1類，可知用類平均法的聚集效果一般。4.1.3離差平方和法聚類 答：用離差平方和法聚類認為較適合的是聚成4類，但庫克群島和西薩摩亞聚成1類，聚集效果尚可。4.1.4 k均值法聚類初始凝聚點：觀測所歸類及與最終凝聚點的距離：聚類匯總：各變量的類內(nèi)均值及標準差：答：用k均值法聚類的話認為較合適的是聚成5類，但庫克群島和西薩摩亞都單成1類，聚集效果一般。4.2對試驗結(jié)果進行合理分析從類平均、離差平方和、k均值三種聚類方