##################################################
####################Random Forests################
##################################################

##########classification################
library(randomForest)
data(iris)
help(iris)
nrow(iris)
head(iris)
set.seed(71)
iris.rf <- randomForest(Species ~ ., data=iris, ntree=2, importance=TRUE,
                        proximity=TRUE)
print(iris.rf)
iris.rf
predict(iris.rf)
importance(iris.rf)
round(importance(iris.rf), 2)
iris.rf$proximity
max(iris.rf$proximity)
min(iris.rf$proximity)
iris.rf$proximity[1:10,]
getTree(randomForest(iris[,-5], iris[,5]), 3, labelVar=TRUE)

#########Regression:########################
data(airquality)
nrow(airquality)
head(airquality)
#regression tree
install.packages("tree")
library(tree)
rt1 = tree(Ozone ~ .,data=airquality)
rt1
plot(rt1); text(rt1)
#random forest
set.seed(131)
newdata<-na.omit(airquality)#construct a new data without the missing value
ozone.rf <- randomForest(Ozone ~ ., data=airquality, mtry=3,
                         importance=TRUE, na.action=na.omit)
print(ozone.rf)
predict(ozone.rf)
## Show "importance" of variables: higher value mean more important:
round(importance(ozone.rf), 2)

#calculate mse 
yhat<-ozone.rf$predicted
yhatnoob<-predict(ozone.rf,airquality[,-1])
mse<-sum(res^2)/nrow(newdata)
#calculate r^2
ssto<-sum((newdata$ozone-mean(newdata$ozone))^2)
sse<-sum(res^2)
rsq<-(ssto-sse)/ssto
rsq