统计建模与r软件第八讲-2017.ppt

总体来看，08统计大约有7个小群体； 红色框内为4对关系密切的同学； 有少数特立独行的样本(同学)存在：王慧、何劲松。 h=read.table(datahomework09.txt) y=read.table(name09.txt) z=as.matrix(y) z=t(z) row.names(h)=z dh=dist(h,method=binary) sdh=scale(dh) hc=hclust(dh) shc=hclust(sdh) plot(hc,hang=-1) x=read.table(datahomework092.txt)#好朋友在哪里 x=as.matrix(x);sigma=var(x) invsigma=solve(sigma) n=50;d=matrix(0,nrow=n,ncol=n) y=read.table(name09.txt) z=as.matrix(y) z=t(z);row.names(d)=z for(i in 1:n) {for(j in 1:i) {muij=x[i,]-x[j,] d[i,j]=sqrt(t(muij)%*%invsigma%*%muij)}} #d[i,j]=mahalanobis(x[i,],x[j,],sigma) d=as.dist(d) hc=hclust(d,method=single) plot(hc,hang=-1) ４.　多分类问题的Bayes判别 样本共分k类:X1,X2,…,Xk,相应的先验概率为p1,p2,…,pk,假定所有的错判损失相同,则判别准则为: 1. ∑1=…= ∑k= ∑ 2. ∑1 ≠ …≠ ∑k R程序 …… if (var.equal == TRUE || var.equal == T){ for (i in 1:g){ d2 - mahalanobis(TstX, mu[i,], var(TrnX)) D[i, ] - d2 - 2*log(p[i]) } } else{ for (i in 1:g){ S-var(TrnX[TrnG==i,]) d2 - mahalanobis(TstX, mu[i,], S) D[i,] - d2 - 2*log(p[i])-log(det(S)) } } for (j in 1:nx){ dmin-Inf for (i in 1:g) if (D[i,j]dmin){ dmin-D[i,j]; blong[j]-i } } blong } 例8.4 用Bayes判别对Fisher Iris数据进行分析.假设先验概率相同,均为1.考虑总体协方差阵不同的情况. x=iris[,1:4] g=gl(3,50) distinguish.bayes(x,g) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 blong 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 blong 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 blong 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 2 3 2 3 2 2 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 blong 2 2 3 2 2 2 2 2 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2