here we are using the inbuilt dataset provided by r and applying c5.0 and rpart decision tree algorithms to classify them in to 3 classes.

DecisionTreeIris.r

#Predict flower species(classify)
iris
head(iris)
dim(iris)
names(iris) 
str(iris)
table(iris$Species)

#split train and test
set.seed(1234)
s = sample(150,100)
iris_train = iris[s,]
iris_test = iris[-s,]
dim(iris_train)
names(iris_train)
dim(iris_test)
#################
#C5.0 is a good classification technique but does not 
#provide good plots. For that rpart is good

library(C50)
Model_C50 <-C5.0(iris_train[,-5],iris_train[,5])
Model_C50
summary(Model_C50)
plot(Model_C50)

#Predicting on Train
P1_train=predict(Model_C50,iris_train);P1_train
table(iris_train[,5],Predicted=P1_train)

#Predicting on Test
P1_test = predict(Model_C50,iris_test);P1_test
table(iris_test[,5],Predicted=P1_test)

#################
#rpart

library(rpart)
library(rpart.plot)
library(party)

Model_rpart= rpart(Species~.,data=iris_train, method="class")

#######plotting rpart using ctree of party rpart(Species~.,data=iris_train, method="class")library

plot(Model_rpart,main="Classifcation Tree for Amount",
     margin=0.15,uniform=TRUE)
text(Model_rpart,use.n=T)

####### 
# Another visualization 

plot(ctree(Species~.,data=iris_train))
Model_rpart
summary(Model_rpart)
# plot(Model_rpart)
# rpart.plot(Model_rpart,type=3)
# rpart.plot(Model_rpart)
#Predicting on Train
P1_train_rpart=predict(Model_rpart,iris_train,type="class")
table(iris_train$Species,predicted=P1_train_rpart)

#Predicting on Test
P1_test_rpart=predict(Model_rpart,iris_test,type="class")


# missing value handing in R using C50 and CART
iris_test$Sepal.Width[iris_test$Sepal.Width==3.7] <- NA
dtc50_test <- predict(dtC50,newdata=iris_test, type="class")
cart_test <- predict(Model_rpart,newdata=iris_test, type="class")