现代回归和分类.pptx

现代回归和分类: 算法建模的回归和分类;算法建模的回归和分类;决策树：分类树和回归树 ;例(数据shuttle.txt)library(MASS);shuttle[1:10,];例 (数据shuttle.txt).;rpart.plot(b,type=4,extra=6,faclen=T); rpart.plot(b,type=0,extra=6,faclen=T);;rpart.plot(b,type=1,extra=6,faclen=T); rpart.plot(b,type=2,extra=6,faclen=T);;例10.1 (数据shuttle.txt).;kyphosis data;The kyphosis data frame has 81 rows and 4 columns. representing data on children who have had corrective spinal surgery Kyphosis: a factor with levels absent present indicating if a kyphosis (a type of deformation) was present after the operation. Age: in months Number: the number of vertebrae involved Start: the number of the first (topmost) vertebra operated on. ;library(rpart.plot);par(mfrow=c(1,3), xpd=NA) ;rpart.plot(fit,type=2,extra=6) rpart.plot(fit2,type=2,extra=6);rpart.plot(fit3,type=2,extra=6);par(mfrow=c(1,1));library(rpart);library(rpart);rpart.plot(kyphosis.rp,type=2,extra=6 );library(rpart);例10.2 (例9.5数据iris.txt). ;rpart.plot(b,type=2,extra=6);IRIS; (数据wine.txt) ;w=read.table(f:/adbook/data/wine1.txt,header=T,sep=,) w$Class=factor(w$Class) library(rpart) (fit - rpart(Class ~ Flavanoids+Color.intensity+Proline,data=w)) plot(fit);text(fit);rpart.plot(fit,type=2,extra=6);思考一下：;回归树 ;例 (数据B1.txt). ;对例9.3数据中的变量crim进行对数变换前后和另外两个变量所做的散点图。 B1=read.table(f:/hepbook/data/Boston.txt,header=T);数据B1.txt;;plot(b);text(b);plot(B1$lstat,B1$crim, xlab=lstat, ylab=crim); partition.tree(b, add = TRUE, cex = 1.5);思考一下：;组合方法：adaboost、bagging和随机森林 ;n=10000;p=.51 s=seq(1,n,2) x=NULL for(i in s)x=c(x,(pbinom(floor(i/2),i,p,lower=F))) plot(s,x,type=l) n=10000;p=.49 s=seq(1,n,2) x1=NULL for(i in s)x1=c(x1,(pbinom(floor(i/2),i,p,lower=F))) plot(s,x1,type=l);par(mfrow=c(1,2));plot(s,x,type=l);plot(s,x1,type=l);par(mfrow=c(2,3)); a=seq(.01,1,.01);n=c(5,11,51,101,1001,9999) for(i in 1:6){ plot(a,1-pbinom(floor(n[i]/2),n[i],a), type=l,main=substitute(n==q,list(q=n[i])), ylab=expression(paste(群体投票决策概率, , p[g])),xlab=expression(paste(个体决策概率, , p))) abline(h=0.5,lty=3);abline(v=0