import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score
from sklearn.tree import export_graphviz

########################################################################

# load data
cars = pd.read_csv('auto-mpg.data.txt',header=None, sep='\s+')

# extract power and weight as data matrix X
X = cars.iloc[:, [3,4]].values

# extract origin as target vector y
y = cars.iloc[:, 7].values

# split into training data (80%) and test data (20%)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)

# learn decision tree
tree = DecisionTreeClassifier(criterion='entropy')
tree.fit(X_train,y_train)
y_predicted = tree.predict(X_test)

# compute confusion matrix
print(confusion_matrix(y_true=y_test, y_pred=y_predicted))

# compute accuracy
print(accuracy_score(y_true=y_test, y_pred=y_predicted)) # 0.620253164556962


# Plotting the Decision Tree
export_graphviz(tree, out_file='tree.dot', feature_names=['Power [HP]', 'Weight [lbs]'], class_names=['U.S.A.','Europe','Japan'])
# you need to have Graphviz (graphviz.org) installed to open the generated file
# to generate a PDF from the .dot file, run:  dot -Tpdf tree.dot -o tree.pdf