09/09/2020 2175052_Safir_ML_Lab2

ML Lab Assignment 2

Team Members
Safir Motiwala (2175052)
Rajendra Kelwa (2175048)

Topic : Find-S Algorithm and Candidate Elimination ALgorithm

Building the DataSet

In [14]:

fields = ['State', 'Climate', 'Age', 'Body_Shape', 'Immune_Strength', 'Covid_19'

]
state = ['Maharashtra', 'Karnataka']
Age = ['Young', 'Old']
Immune_Strength = ['Weak', 'High']
Covid_19 = ['Yes', 'No']

import csv
import random

filename = 'Covid-19_Data.csv'
with open(filename, 'w') as csvfile:
csvwriter = csv.writer(csvfile)
csvwriter.writerow(fields)
row = [random.choice(state), 'Sunny', random.choice(Age), 'Under-weight', ra
ndom.choice(Immune_Strength), 'Yes']
csvwriter.writerow(row)
for i in range(2):
row = [random.choice(state), 'Sunny', random.choice(Age), 'Under-weight'
, random.choice(Immune_Strength), 'No']
csvwriter.writerow(row)
for i in range(0, 390):
row = [random.choice(state), 'Sunny', random.choice(Age), 'Under-weight'
, random.choice(Immune_Strength), random.choice(Covid_19)]
csvwriter.writerow(row)
for i in range(0,10):
row = [random.choice(state), 'Winter', random.choice(Age), 'Over-weight'
, random.choice(Immune_Strength), random.choice(Covid_19)]
csvwriter.writerow(row)

Importing the Dataset and Preprocessing

localhost:8889/nbconvert/html/Desktop/ML_Lab/2175052_Safir_ML_Lab2.ipynb?download=false 1/9
09/09/2020 2175052_Safir_ML_Lab2

In [2]:

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

dataset = pd.read_csv('Covid-19_Data.csv')

result = {'Yes':1, 'No':0}

dataset['Covid_19'] = dataset['Covid_19'].map(result)

X = dataset.iloc[:, 0:5].values
y = dataset.iloc[:, -1].values

from sklearn.model_selection import KFold

kf = KFold(n_splits=10)
#print(kf.get_n_splits(X))

for train_index, test_index in kf.split(X,y):

X_train, X_test = X[train_index], X[test_index]
y_train, y_test = y[train_index], y[test_index]

In [3]:

print(dataset.head())

State Climate Age Body_Shape Immune_Strength Covid_1

9
0 Maharashtra Sunny Young Under-weight Weak
1
1 Maharashtra Sunny Young Under-weight Weak
0
2 Karnataka Sunny Old Under-weight Weak
0
3 Maharashtra Sunny Old Under-weight High
1
4 Maharashtra Sunny Young Under-weight Weak
0

Writing the Code for Find-S Algorithm

localhost:8889/nbconvert/html/Desktop/ML_Lab/2175052_Safir_ML_Lab2.ipynb?download=false 2/9
09/09/2020 2175052_Safir_ML_Lab2

In [4]:

class FindS:
def __init__(self):
self.Xtrain = ""
self.ytrain = ""
self.Xtest = ""
self.ytest = ""
self.specific_hypothesis = []

def fit(self, X, y):

count = 0
self.Xtrain, self.ytrain = X, y
for i, val in enumerate(y):
if val==1:
S_hypothesis = list(X[0].copy())
print("Initial : ", S_hypothesis)
break
for i, val in enumerate(X):
if y[i]==1:
count+=1
for x in range(len(S_hypothesis)):
if val[x] == S_hypothesis[x]:
pass
else:
S_hypothesis[x] = '?'
print("Total Positive Records : ", count)
self.specific_hypothesis = S_hypothesis
return S_hypothesis

def predict(self, X):

y = np.array([0 for i in range(len(X))])
self.Xtest = X
for i, val in enumerate(X):
val = list(val)
check = 0
for x in range(len(val)):
if val[x] == self.specific_hypothesis[x]:
check+=1
else:
pass
if check>0:
y[i] = 1
else:
y[i] = 0
self.ytest = y
return y

Fitting the Training Set to the Model

In [5]:

from FindSAlgorithm import FindS

fs = FindS()
S_hypothesis = fs.fit(X_train, y_train)

Initial : ['Maharashtra', 'Sunny', 'Young', 'Under-weight', 'Weak']

Total Positive Records : 180

localhost:8889/nbconvert/html/Desktop/ML_Lab/2175052_Safir_ML_Lab2.ipynb?download=false 3/9
09/09/2020 2175052_Safir_ML_Lab2

In [6]:

print("Specific Hypothesis : ", S_hypothesis)

Specific Hypothesis : ['?', 'Sunny', '?', 'Under-weight', '?']

Predicting the Test Set Results

In [7]:

y_pred = fs.predict(X_test)
y_pred1 = fs.predict(X_train)

In [8]:

print(y_pred)

[1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
0 0 0
0 0 0]

Finding the Average Precision Score

In [9]:

from sklearn.metrics import accuracy_score, average_precision_score, f1_score

Avg1 = average_precision_score(y_train, y_pred1, average='micro', pos_label=1)

Avg2 = average_precision_score(y_test, y_pred, average='micro', pos_label=1)

print("Average Precision Score (Training Data) : ", Avg1)

print("Average Precision Score (Test Data) : ", Avg2)

Average Precision Score (Training Data) : 0.5

Average Precision Score (Test Data) : 0.33809523809523806

Finding the F-1 Score

In [10]:

FS1 = f1_score(y_train, y_pred1, average='binary', pos_label=1)

FS2 = f1_score(y_test, y_pred, average='binary', pos_label=1)

print("F-1 Score (Training Data) : ", FS1)

print("F-1 Score (Test Data) : ", FS2)

F-1 Score (Training Data) : 0.6666666666666666

F-1 Score (Test Data) : 0.4545454545454545

Plotting Area Under the Curve

localhost:8889/nbconvert/html/Desktop/ML_Lab/2175052_Safir_ML_Lab2.ipynb?download=false 4/9
09/09/2020 2175052_Safir_ML_Lab2

In [11]:

from sklearn import metrics

fpr, tpr, thresholds = metrics.roc_curve(y_test, y_pred, pos_label=1)
auc = metrics.auc(fpr, tpr)
print("Area under the Curve : ", auc)

Area under the Curve : 0.4725274725274725

In [12]:

from sklearn.metrics import roc_auc_score, roc_curve

ns_probs = [0 for _ in range(len(y_test))]
lr_probs = y_pred

ns_auc = roc_auc_score(y_test, ns_probs)

lr_auc = roc_auc_score(y_test, lr_probs)

print('No Skill: ROC AUC=%.3f' % (ns_auc))

print('Logistic: ROC AUC=%.3f' % (lr_auc))

ns_fpr, ns_tpr, _ = roc_curve(y_test, ns_probs)

lr_fpr, lr_tpr, _ = roc_curve(y_test, lr_probs)
plt.plot(ns_fpr, ns_tpr, linestyle='--', label='No Skill')
plt.plot(lr_fpr, lr_tpr, marker='.', label='Logistic')
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.legend()
plt.show()

No Skill: ROC AUC=0.500

Logistic: ROC AUC=0.473

Writing the Code for Candidate Elimination Algorithm

localhost:8889/nbconvert/html/Desktop/ML_Lab/2175052_Safir_ML_Lab2.ipynb?download=false 5/9
09/09/2020 2175052_Safir_ML_Lab2

In [13]:

class Candidate_Elimination:
def __init__(self):
self.Xtrain = ""
self.ytrain = ""
self.Xtest = ""
self.ytest = ""
self.specific_hypothesis = []
self.general_hypothesis = []
self.version_space = []

def fit(self, X, y):

count, count1 = 0, 0
self.Xtrain, self.ytrain = X, y
for i, val in enumerate(y):
if val==1:
S_hypothesis = list(X[0].copy())
G_hypothesis = [['?' for _ in range(len(S_hypothesis))]]
print("Initial Specific Hypothesis : ", S_hypothesis)
print("Initial General Hypothesis : ", G_hypothesis)
break
for i, val in enumerate(X):
if y[i]==1:
count+=1
for x in range(len(S_hypothesis)):
#print(G_hypothesis)
if val[x] == S_hypothesis[x]:
pass
else:
S_hypothesis[x] = '?'
temp = ['?' for _ in range(len(S_hypothesis))]
temp[x] = val[x]
#print(temp)
if temp not in G_hypothesis:
pass
else:
G_hypothesis.remove(temp)
#print("Removed")
elif y[i]==0:
count1+=1
temp = []
for x in range(len(S_hypothesis)):
if val[x] != S_hypothesis[x] and S_hypothesis[x] != '?':
temp = ['?' for _ in range(len(S_hypothesis))]
temp[x] = S_hypothesis[x]
if temp not in G_hypothesis:
G_hypothesis.append(temp)
temp = []

if len(G_hypothesis)==1:
v1 = [S_hypothesis]
else:
v1 = []
for i in G_hypothesis:
for j in range(len(i)):
if i[j] != '?':
h = i[j]
for z in range(len(S_hypothesis)):
if S_hypothesis[z] != '?':
temp = ['?' for _ in range(len(S_hypothesis))]
localhost:8889/nbconvert/html/Desktop/ML_Lab/2175052_Safir_ML_Lab2.ipynb?download=false 6/9
09/09/2020 2175052_Safir_ML_Lab2

temp[z] = S_hypothesis[z]
temp[j] = i[j]
v1.append(temp)

print('Final Specific Hypothesis : ', S_hypothesis)

print('Final General Hypothesis : ', G_hypothesis)
print('Final Version Space : ', v1)
print(count, " ", count1)
self.specific_hypothesis, self.general_hypothesis = S_hypothesis, G_hypo
thesis
self.version_space = v1
return 1

def predict(self, X):

y = np.array([0 for i in range(len(X))])
self.Xtest = X
for i, val in enumerate(X):
val = list(val)
check = 0
for x in range(len(val)):
for z in self.version_space:
if val[x] == z[x]:
check+=1
else:
pass
if check>0:
y[i] = 1
else:
y[i] = 0
self.ytest = y
return y

Fitting the Training Set to the Model

In [15]:

from CandidateEliminationAlgorithm import Candidate_Elimination

ce = Candidate_Elimination()
ce.fit(X_train, y_train)

Initial Specific Hypothesis : ['Maharashtra', 'Sunny', 'Young', 'Un

der-weight', 'Weak']
Initial General Hypothesis : [['?', '?', '?', '?', '?']]
Final Specific Hypothesis : ['?', 'Sunny', '?', 'Under-weight',
'?']
Final General Hypothesis : [['?', '?', '?', '?', '?']]
Final Version Space : [['?', 'Sunny', '?', 'Under-weight', '?']]
180 180

Out[15]:

Predicting the Test Set Results

localhost:8889/nbconvert/html/Desktop/ML_Lab/2175052_Safir_ML_Lab2.ipynb?download=false 7/9
09/09/2020 2175052_Safir_ML_Lab2

In [16]:

y_pred = ce.predict(X_test)
y_pred1 = ce.predict(X_train)

In [17]:

print(y_pred)

[1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
0 0 0
0 0 0]

Finding the Average Precision Score

In [18]:

from sklearn.metrics import accuracy_score, average_precision_score, f1_score

Avg1 = average_precision_score(y_train, y_pred1, average='micro', pos_label=1)

Avg2 = average_precision_score(y_test, y_pred, average='micro', pos_label=1)

print("Average Precision Score (Training Data) : ", Avg1)

print("Average Precision Score (Test Data) : ", Avg2)

Average Precision Score (Training Data) : 0.5

Average Precision Score (Test Data) : 0.33809523809523806

Finding the F-1 Score

In [19]:

FS1 = f1_score(y_train, y_pred1, average='binary', pos_label=1)

FS2 = f1_score(y_test, y_pred, average='binary', pos_label=1)

print("F-1 Score (Training Data) : ", FS1)

print("F-1 Score (Test Data) : ", FS2)

F-1 Score (Training Data) : 0.6666666666666666

F-1 Score (Test Data) : 0.4545454545454545

Plotting the Area Under the Curve

In [20]:

from sklearn import metrics

fpr, tpr, thresholds = metrics.roc_curve(y_test, y_pred, pos_label=1)
auc = metrics.auc(fpr, tpr)
print("Area under the Curve : ", auc)

Area under the Curve : 0.4725274725274725

localhost:8889/nbconvert/html/Desktop/ML_Lab/2175052_Safir_ML_Lab2.ipynb?download=false 8/9
09/09/2020 2175052_Safir_ML_Lab2

In [21]:

from sklearn.metrics import roc_auc_score, roc_curve

ns_probs = [0 for _ in range(len(y_test))]
lr_probs = y_pred

ns_auc = roc_auc_score(y_test, ns_probs)

lr_auc = roc_auc_score(y_test, lr_probs)

print('No Skill: ROC AUC=%.3f' % (ns_auc))

print('Logistic: ROC AUC=%.3f' % (lr_auc))

ns_fpr, ns_tpr, _ = roc_curve(y_test, ns_probs)

lr_fpr, lr_tpr, _ = roc_curve(y_test, lr_probs)

plt.plot(ns_fpr, ns_tpr, linestyle='--', label='No Skill')

plt.plot(lr_fpr, lr_tpr, marker='.', label='Logistic')
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.legend()
plt.show()

No Skill: ROC AUC=0.500

Logistic: ROC AUC=0.473

In [ ]:

localhost:8889/nbconvert/html/Desktop/ML_Lab/2175052_Safir_ML_Lab2.ipynb?download=false 9/9