實驗報告材料 聚類分析報告

《實驗報告材料 聚類分析報告》由會員分享，可在線閱讀，更多相關《實驗報告材料 聚類分析報告（28頁珍藏版）》請在裝配圖網上搜索。

1、word 實驗報告 聚類分析 實驗原理：K均值聚類、中心點聚類、系統(tǒng)聚類和EM算法聚類分析技術。 實驗題目：用鳶尾花的數(shù)據(jù)集，進行聚類挖掘分析。 實驗要求：探索鳶尾花數(shù)據(jù)的基本特征，利用不同的聚類挖掘方法，獲得基本結論并簡明解釋。 實驗題目--分析報告：data(iris) > rm(list=ls()) > gc() used (Mb) gc trigger (Mb) max used (Mb) Ncells 431730 23.1 929718 49.7 607591 32.5 Vcells 787605 6.1 8388608 6

2、4.0 1592403 12.2 > data(iris) > data<-iris > head(data) Sepal.Length Sepal.Width Petal.Length Petal.Width Species 1 5.1 3.5 1.4 0.2 setosa 2 4.9 3.0 1.4 0.2 setosa 3 4.7 3.2 1.3 0.2 s

3、etosa 4 4.6 3.1 1.5 0.2 setosa 5 5.0 3.6 1.4 0.2 setosa 6 5.4 3.9 1.7 0.4 setosa #Kmean聚類分析 > newiris <- iris > newiris$Species <- NULL > (kc <- kmeans(newiris, 3)) K-means clustering w

4、ith 3 clusters of sizes 62, 50, 38 Cluster means: Sepal.Length Sepal.Width Petal.Length Petal.Width 1 5.901613 2.748387 4.393548 1.433871 2 5.006000 3.428000 1.462000 0.246000 3 6.850000 3.073684 5.742105 2.071053 Clustering vector: [1] 2 2 2

5、2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 [41] 2 2 2 2 2 2 2 2 2 2 1 1 3 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 1 1 [81] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 1 3 3 3 3 1 3 3 3 3 3 3 1 1 3 3 3 3 1 [121] 3 1 3 1 3 3 1 1 3 3 3 3 3 1 3 3 3 3 1 3 3 3 1

6、 3 3 3 1 3 3 1 Within cluster sum of squares by cluster: [1] 39.82097 15.15100 23.87947 (between_SS / total_SS = 88.4 %) Available ponents: [1] "cluster" "centers" "totss" "withinss" "tot.withinss" [6] "betweenss" "size" "iter" "ifault" >

7、table(iris$Species, kc$cluster) 1 2 3 setosa 0 50 0 versicolor 48 0 2 virginica 14 0 36 > plot(newiris[c("Sepal.Length", "Sepal.Width")], col = kc$cluster) > points(kc$centers[,c("Sepal.Length", "Sepal.Width")], col = 1:3, pch = 8, cex=2) #K-Mediods

8、 進行聚類分析 > install.packages("cluster") > library(cluster) > iris.pam<-pam(iris,3) > table(iris$Species,iris.pam$clustering) 1 2 3 setosa 50 0 0 versicolor 0 3 47 virginica 0 49 1 > layout(matrix(c(1,2),1,2)) > plot(iris.pam) > layout(matrix(1)) #hc

9、> iris.hc <- hclust( dist(iris[,1:4])) > plot( iris.hc, hang = -1) > plclust( iris.hc, labels = FALSE, hang = -1) > re <- rect.hclust(iris.hc, k = 3) > iris.id <- cutree(iris.hc, 3) #利用剪枝函數(shù)cutree()參數(shù)h控制輸出height=18時的系譜類別 > sapply(unique(iris.id), + function(g)iris$Species[iris.id==g])

10、 [[1]] [1] setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa [12] setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa [23] setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa setosa [34] setosa setosa setosa setosa setosa se

11、tosa setosa setosa setosa setosa setosa [45] setosa setosa setosa setosa setosa setosa Levels: setosa versicolor virginica [[2]] [1] versicolor versicolor versicolor versicolor versicolor versicolor versicolor [8] versicolor versicolor versicolor versicolor versicolor versicolor versicolor

12、 [15] versicolor versicolor versicolor versicolor versicolor versicolor versicolor [22] versicolor versicolor virginica virginica virginica virginica virginica [29] virginica virginica virginica virginica virginica virginica virginica [36] virginica virginica virginica virginica v

13、irginica virginica virginica [43] virginica virginica virginica virginica virginica virginica virginica [50] virginica virginica virginica virginica virginica virginica virginica [57] virginica virginica virginica virginica virginica virginica virginica [64] virginica vi

14、rginica virginica virginica virginica virginica virginica [71] virginica virginica Levels: setosa versicolor virginica [[3]] [1] versicolor versicolor versicolor versicolor versicolor versicolor versicolor [8] versicolor versicolor versicolor versicolor versicolor versicolor versico

15、lor [15] versicolor versicolor versicolor versicolor versicolor versicolor versicolor [22] versicolor versicolor versicolor versicolor versicolor versicolor virginica Levels: setosa versicolor virginica > plot(iris.hc) > rect.hclust(iris.hc,k=4,border="light grey")#用淺灰色矩形框出4分類聚類結果 > rect.hclu

16、st(iris.hc,k=3,border="dark grey")#用淺灰色矩形框出3分類聚類結果 > rect.hclust(iris.hc,k=7,which=c(2,6),border="dark grey") # DBSCAN #基于密度的聚類 > install.packages("fpc") > library(fpc) > ds1=dbscan(iris[,1:4],eps=1,MinPts=5)#半徑參數(shù)為1，密度閾值為5 > ds1 dbscan Pts=150 MinPts=5 eps=1 1 2 border 0 1 s

17、eed 50 99 total 50 100 > ds2=dbscan(iris[,1:4],eps=4,MinPts=5) > ds3=dbscan(iris[,1:4],eps=4,MinPts=2) > ds4=dbscan(iris[,1:4],eps=8,MinPts=2) > par(mfcol=c(2,2)) > plot(ds1,iris[,1:4],main="1: MinPts=5 eps=1") > plot(ds3,iris[,1:4],main="3: MinPts=2 eps=4") > plot(ds2,iris[,1:4],main="2

18、: MinPts=5 eps=4") > plot(ds4,iris[,1:4],main="4: MinPts=2 eps=8") > d=dist(iris[,1:4])#計算數(shù)據(jù)集的距離矩陣d > max(d);min(d)#計算數(shù)據(jù)集樣本的距離的最值 [1] 7.085196 [1] 0 > install.packages("ggplot2") > library(ggplot2) > interval=cut_interval(d,30) > table(interval) interval [0,0.236] (0.236,0.472] (0.4

19、72,0.709] (0.709,0.945] (0.945,1.18] (1.18,1.42] 88 585 876 891 831 688 (1.42,1.65] (1.65,1.89] (1.89,2.13] (2.13,2.36] (2.36,2.6] (2.6,2.83] 543 369 379 339 335

20、 406 (2.83,3.07] (3.07,3.31] (3.31,3.54] (3.54,3.78] (3.78,4.01] (4.01,4.25] 458 459 465 480 468 505 (4.25,4.49] (4.49,4.72] (4.72,4.96] (4.96,5.2] (5.2,5.43] (5.43,5.67] 349 385

21、 321 291 187 (5.67,5.9] (5.9,6.14] (6.14,6.38] (6.38,6.61] (6.61,6.85] (6.85,7.09] 97 92 78 50 18 4 > which.max(table(interval)) (0.709,0.945] 4 > for(i in 3:

22、5) + { for(j in 1:10) + { ds=dbscan(iris[,1:4],eps=i,MinPts=j) + print(ds) + } + } dbscan Pts=150 MinPts=1 eps=3 1 seed 150 total 150 dbscan Pts=150 MinPts=2 eps=3 1 seed 150 total 150 dbscan Pts=150 MinPts=3 eps=3 1 seed 150 total 150 dbscan Pt

23、s=150 MinPts=4 eps=3 1 seed 150 total 150 dbscan Pts=150 MinPts=5 eps=3 1 seed 150 total 150 dbscan Pts=150 MinPts=6 eps=3 1 seed 150 total 150 dbscan Pts=150 MinPts=7 eps=3 1 seed 150 total 150 dbscan Pts=150 MinPts=8 eps=3 1 seed 150

24、total 150 dbscan Pts=150 MinPts=9 eps=3 1 seed 150 total 150 dbscan Pts=150 MinPts=10 eps=3 1 seed 150 total 150 dbscan Pts=150 MinPts=1 eps=4 1 seed 150 total 150 dbscan Pts=150 MinPts=2 eps=4 1 seed 150 total 150 dbscan Pts=150 MinPts=3 eps=4

25、 1 seed 150 total 150 dbscan Pts=150 MinPts=4 eps=4 1 seed 150 total 150 dbscan Pts=150 MinPts=5 eps=4 1 seed 150 total 150 dbscan Pts=150 MinPts=6 eps=4 1 seed 150 total 150 dbscan Pts=150 MinPts=7 eps=4 1 seed 150 total 150 dbscan Pts=1

26、50 MinPts=8 eps=4 1 seed 150 total 150 dbscan Pts=150 MinPts=9 eps=4 1 seed 150 total 150 dbscan Pts=150 MinPts=10 eps=4 1 seed 150 total 150 dbscan Pts=150 MinPts=1 eps=5 1 seed 150 total 150 dbscan Pts=150 MinPts=2 eps=5 1 seed 150 to

27、tal 150 dbscan Pts=150 MinPts=3 eps=5 1 seed 150 total 150 dbscan Pts=150 MinPts=4 eps=5 1 seed 150 total 150 dbscan Pts=150 MinPts=5 eps=5 1 seed 150 total 150 dbscan Pts=150 MinPts=6 eps=5 1 seed 150 total 150 dbscan Pts=150 MinPts=7 eps=5

28、 1 seed 150 total 150 dbscan Pts=150 MinPts=8 eps=5 1 seed 150 total 150 dbscan Pts=150 MinPts=9 eps=5 1 seed 150 total 150 dbscan Pts=150 MinPts=10 eps=5 1 seed 150 total 150 #30次dbscan的聚類結果 > ds5=dbscan(iris[,1:4],eps=3,MinPts=2) > ds6=dbscan(iris[,1

29、:4],eps=4,MinPts=5) > ds7=dbscan(iris[,1:4],eps=5,MinPts=9) > par(mfcol=c(1,3)) > plot(ds5,iris[,1:4],main="1: MinPts=2 eps=3") > plot(ds6,iris[,1:4],main="3: MinPts=5 eps=4") > plot(ds7,iris[,1:4],main="2: MinPts=9 eps=5") # EM 期望最大化聚類 > install.packages("mclust") > library(mclust) > fit

30、_EM=Mclust(iris[,1:4]) fitting ... |===========================================================================| 100% > summary(fit_EM) ---------------------------------------------------- Gaussian finite mixture model fitted by EM algorithm ------------------------------------------------

31、---- Mclust VEV (ellipsoidal, equal shape) model with 2 ponents: log.likelihood n df BIC ICL -215.726 150 26 -561.7285 -561.7289 Clustering table: 1 2 50 100 > summary(fit_EM,parameters=TRUE) ---------------------------------------------------- Gaussian f

32、inite mixture model fitted by EM algorithm ---------------------------------------------------- Mclust VEV (ellipsoidal, equal shape) model with 2 ponents: log.likelihood n df BIC ICL -215.726 150 26 -561.7285 -561.7289 Clustering table: 1 2 50 100 Mix

33、ing probabilities: 1 2 0.3333319 0.6666681 Means: [,1] [,2] Sepal.Length 5.0060022 6.261996 Sepal.Width 3.4280049 2.871999 Petal.Length 1.4620007 4.905992 Petal.Width 0.2459998 1.675997 Variances: [,,1] Sepal.Length Sepal.Width Pe

34、tal.Length Petal.Width Sepal.Length 0.15065114 0.13080115 0.02084463 0.01309107 Sepal.Width 0.13080115 0.17604529 0.01603245 0.01221458 Petal.Length 0.02084463 0.01603245 0.02808260 0.00601568 Petal.Width 0.01309107 0.01221458 0.00601568 0.01042365 [,,2]

35、Sepal.Length Sepal.Width Petal.Length Petal.Width Sepal.Length 0.4000438 0.10865444 0.3994018 0.14368256 Sepal.Width 0.1086544 0.10928077 0.1238904 0.07284384 Petal.Length 0.3994018 0.12389040 0.6109024 0.25738990 Petal.Width 0.1436826 0.07284384 0.2573899 0.168

36、08182 > plot(fit_EM)#對EM聚類結果作圖 Model-based clustering plots: 1: BIC 2: classification 3: uncertainty 4: density Selection:（下面顯示選項） #選1 #選2 #選3 #選4 Selection: 0 > iris_BIC=mclustBIC(iris[,1:4]) fitting ... |==================================================================

37、=========| 100% > iris_BICsum=summary(iris_BIC,data=iris[,1:4]) > iris_BICsum #獲取數(shù)1據(jù)集iris在各模型和類別數(shù)下的BIC值 Best BIC values: VEV,2 VEV,3 VVV,2 BIC -561.7285 -562.5522369 -574.01783 BIC diff 0.0000 -0.8237748 -12.28937 Classification table for model (VEV,2):

38、 1 2 50 100 > iris_BIC Bayesian Information Criterion (BIC): EII VII EEI VEI EVI VVI EEE 1 -1804.0854 -1804.0854 -1522.1202 -1522.1202 -1522.1202 -1522.1202 -829.9782 2 -1123.4117 -1012.2352 -1042.9679 -956.2823 -1007.3082 -857.5515

39、 -688.0972 3 -878.7650 -853.8144 -813.0504 -779.1566 -797.8342 -744.6382 -632.9647 4 -893.6140 -812.6048 -827.4036 -748.4529 -837.5452 -751.0198 -646.0258 5 -782.6441 -742.6083 -741.9185 -688.3463 -766.8158 -711.4502 -604.8131 6 -715.7136 -705.7811 -693.7908 -676.1697 -774

40、.0673 -707.2901 -609.8543 7 -731.8821 -698.5413 -713.1823 -680.7377 -813.5220 -766.6500 -632.4947 8 -725.0805 -701.4806 -691.4133 -679.4640 -740.4068 -764.1969 -639.2640 9 -694.5205 -700.0276 -696.2607 -702.0143 -767.8044 -755.8290 -653.0878 EVE VEE VVE

41、 EEV VEV EVV VVV 1 -829.9782 -829.9782 -829.9782 -829.9782 -829.9782 -829.9782 -829.9782 2 -657.2263 -656.3270 -605.1841 -644.5997 -561.7285 -658.3306 -574.0178 3 -666.5491 -605.3982 -636.4259 -644.7810 -562.5522 -656.0359 -580.8396 4 -705.5435 -604.8371 -639.7078 -699.8684 -

42、602.0104 -725.2925 -630.6000 5 -723.7199 NA -632.2056 -652.2959 -634.2890 NA -676.6061 6 -661.9497 -609.5584 -664.8224 -664.4537 -679.5116 NA -754.7938 7 -699.5102 NA -690.6108 -709.9530 -704.7699 -809.8276 -806.9277 8 -700.4277 -654.8237 -709.9392 -735.4463 -712.8788

43、 -831.7520 -830.6373 9 -729.6651 NA -734.2997 -758.9348 -748.8237 -882.4391 -883.6931 Top 3 models based on the BIC criterion: VEV,2 VEV,3 VVV,2 -561.7285 -562.5522 -574.0178 > par(mfcol=c(1,1)) > plot(iris_BIC,G=1:7,col="yellow") > mclust2Dplot(iris[,1:2], +

44、 classification=iris_BICsum$classification, + parameters=iris_BICsum$parameters,col="yellow") > iris_Dens=densityMclust(iris[,1:2])# 對每一個樣本進行密度估計 fitting ... |===========================================================================| 100% > iris_Dens 'densityMclust' model objec

45、t: (VEV,2) Available ponents: [1] "call" "data" "modelName" "n" [5] "d" "G" "BIC" "bic" [9] "loglik" "df" "hypvol" "parameters" [13] "z" "classification" "un

46、certainty" "density" > plot(iris_Dens,iris[,1:2],col="yellow",nlevels=55) ##輸入1或2 Model-based density estimation plots: 1: BIC 2: density Selection:（下面顯示選項） #選1 #選2 Selection: 0 > plot(iris_Dens,type = "persp",col = grey(0.8)) Model-based density estimation plots: 1: BIC 2: density Selection:（下面顯示選項） #選1 #選2 Selection: 0 28 / 28