hospital = read.csv("hospital.csv")
table(hospital$STATE)
sum(table(hospital$STATE)[1:12])
names(table(hospital$STATE)[1:12])-> su 
i = !is.na(match(hospital$STATE,su))
hosp = hospital[i,-c(1:4,10,11)]
y = log(1+hospital[i,13])
library(rpart)
htree = rpart(y~., data=hosp)
par(cex=0.5); plot(htree) ; text(htree)
plot(hosp[,12],y)
u=rpart(y~.,data=hosp,control=rpart.control(cp=.005))
u
plot(u)
u=rpart(y~.,data=hosp,control=rpart.control(cp=.0001))
plot(u)
u
hosp[1,]
u=rpart(y~.,data=hosp[,c(-12,-7)],control=rpart.control(cp=.0001))
u
u=rpart(y~.,data=hosp[,c(-12,-7)],control=rpart.control(cp=.001))
plot(u)
plot(silhouette(pam(ruspini, k=4)), main = paste("k = ",4), do.n.k=FALSE)
library(cluster)
pam(hout[,2:5], k=17)-> hpam
table(hpam$cluster)
table(hpam$cluster, hout[,6])