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

# # Import

# In[1]:


import pandas as pd
import os
from sklearn.cross_validation import train_test_split
from sklearn.metrics import accuracy_score
from sklearn.svm import SVC
from sklearn.metrics import confusion_matrix

print pd.__version__


# In[2]:


# reading data into a dataframe
DATA_DIR = 'data'

df = pd.read_csv(
                    os.path.abspath(os.path.join(DATA_DIR, 'day9/banknote_authentication.csv'))
                )
df.head(5)


# In[3]:


# (rows, columns)
df.shape


# In[4]:


# look for NaN values
df.isnull().sum()


# In[5]:


# let's figure out the distribution of target variable
# there are many options to handle this imbalance; of which one is to add false data to class 1; secondly we can 
# delete records from 0 to make it equal to 1
# we will leave it for now
df['target'].value_counts()


# In[6]:


X = df.iloc[:, :-1].values
Y = df.iloc[:, -1].values


# # Split

# In[7]:


# ideal practice is to use test as 20% - 30% of training data
# defined by test_size in train_test_split()
# random_state is required to avoid sequential biasness in the data distribution
def data_split(X, Y):
    X_train, X_test, Y_train, Y_test = train_test_split( X, Y, test_size=0.3, random_state = 10)
    return X_train, X_test, Y_train, Y_test

X_train, X_test, Y_train, Y_test = data_split(X, Y)


# In[8]:


X_train.shape, X_test.shape


# # Model Define

# In[9]:


class SVMModel:
    
    def __init__(self):
        self.classifier = SVC()

    def train(self, X_train, Y_train):
        model = self.classifier.fit(X_train, Y_train)
        return model
    
    def predict(self, model, X_test):
        return model.predict(X_test)
    
    def evaluate(self, Y_test, Y_pred, measure):
        if measure=='matrix':
            cm = confusion_matrix(Y_test, Y_pred, labels=[0, 1])
            return cm
        elif measure=='accuracy':
            return accuracy_score(Y_test, Y_pred)*100
        else: return None


# # Training

# In[10]:


# train the model
svm = SVMModel()
model = svm.train(X_train, Y_train)
predictions = svm.predict(model, X_test)


# In[12]:


# evaluating the model
print svm.evaluate(Y_test, predictions, 'matrix')
print 
print svm.evaluate(Y_test, predictions, 'accuracy')