#!/usr/bin/env python
# coding: utf-8

# ## DecisionTree Classification 
# 
# [https://github.com/arundhaj](https://github.com/arundhaj)
# 
# using vertebrate data set

# In[1]:


get_ipython().run_line_magic('matplotlib', 'inline')

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt


# In[2]:


# Read the CSV file
data = pd.read_csv('data/vertebrate.csv')


# In[3]:


# List of all classes
data['Class Label']


# In[4]:


# List of unique classes
data['Class Label'].unique()


# In[5]:


# Number of entries for each unique classes
class_group = data.groupby('Class Label').apply(lambda x: len(x))
class_group


# In[6]:


# Plot bar chart based on Class Label
class_group.plot(kind='bar', grid=False)


# In[7]:


from sklearn.preprocessing import LabelEncoder
from sklearn.feature_extraction import DictVectorizer

cols_to_retain = ['Body Temperature', 'Skin Cover', 'Gives Birth', 'Aquatic Creature', 'Aerial Creature', 'Has Legs', 'Hibernates']

X_feature = data[cols_to_retain]
X_dict = X_feature.T.to_dict().values()

# turn list of dicts into a numpy array
vect = DictVectorizer(sparse=False)
X_vector = vect.fit_transform(X_dict)

# print the features
# vect.get_feature_names()

# 0 to 14 is train set
X_Train = X_vector[:-1]
# 15th is test set
X_Test = X_vector[-1:] 

# Used to vectorize the class label
le = LabelEncoder()
y_train = le.fit_transform(data['Class Label'][:-1])


# In[8]:


from sklearn import tree

clf = tree.DecisionTreeClassifier(criterion='entropy')
clf = clf.fit(X_Train,y_train)


# In[9]:


# Predict the test data, not seen earlier
le.inverse_transform(clf.predict(X_Test))


# In[10]:


# prediction with the same training set
Train_predict = clf.predict(X_Train)


# In[11]:


# The model predicted the training set correctly
(Train_predict == y_train).all()


# In[12]:


# Metrics related to the DecisionTreeClassifier
from sklearn.metrics import accuracy_score, classification_report

print 'Accuracy is:', accuracy_score(y_train, Train_predict)
print classification_report(y_train, Train_predict)


# In[13]:


'''
import pydot
import pyparsing

import StringIO
dot_data = StringIO.StringIO() 
tree.export_graphviz(clf, out_file=dot_data) 
graph = pydot.graph_from_dot_data(dot_data.getvalue()) 
graph.write_png('data/vertebrate/tree.png') 
from IPython.core.display import Image 
Image(filename='data/vertebrate/tree.png')
'''


# In[ ]: