Assg6/7

import numpy as np
import pandas as pd

data = pd.read_csv("Dataset/sales.csv")
data

data.describe()

#data.apply(pd.value_counts)
data['Buys'].value_counts()

from sklearn.preprocessing import LabelEncoder
le=LabelEncoder();
#data=data.apply(le.fit_transform)
x=data.iloc[:,:-1] #-1 means don't take last column
x=x.apply(le.fit_transform)
#find label with their encoded value
print("Age with encodd value :",list( zip(data.iloc[:,0], x.iloc[:,0])))
print("\nIncome with encoded value :",list( zip(data.iloc[:,1], x.iloc[:,1])))
print("\nGender with encoded value :",list( zip(data.iloc[:,2], x.iloc[:,2])))
print("\nmaritialStatus with encoded value :",list( zip(data.iloc[:,3], x.iloc[:,3])))

y=data.iloc[:,-1]

from sklearn.tree import DecisionTreeClassifier
classifier=DecisionTreeClassifier(criterion='entropy')
classifier.fit(x,y)

#Predict value for the given Expression
#[Age < 21, Income = Low,Gender = Female, Marital Status = Married]
test_x=np.array([2,2,0,0])
pred_y=classifier.predict([test_x])
print("Predicted class for input [Age > 35, Income = Medium,Gender = Female, Marital Status = Married]\n", test_x," is ",pred_y[0])

#method to generate graph p.s. needs dot utility installed in os
from sklearn.tree import export_graphviz
from IPython.display import Image
export_graphviz(classifier,out_file="6th7th.dot",feature_names=x.columns,class_names=["No","Yes"])
#you need to install graphviz in fedora(IN LAB) for running below dor command
#yum install graphviz
#then go to terminal and cd to directory where you are saving jupyter notebook
# and execute below command
#    dot -Tpng data.dot -o tree.png
# !dot -Tpng data.dot -o tree.png
# Image("tree.png")