Lab Manual: Department of Computer Science and Engineering

DEPARTMENT OF COMPUTER SCIENCE
AND ENGINEERING
LAB MANUAL
B.Tech. VI Semester
MACHINE LEARNING LAB

6CS4-22
Modern Institute of Technology & Research Centre,

Alwar, Rajasthan
Affiliated to Bikaner Technical University, Bikaner
6CS4-22: Machine Learning Lab
TABLE OF CONTENT
S.No Content Page No
Lab Instruction i
BTU Syllabus ii
Lab Introduction iii
Implement and demonstrate the FIND-S algorithm for finding the
1 most specific hypothesis based on a given set of training data 1-2
samples.
For a given set of training data examples, implement and
2 demonstrate the Candidate-Elimination algorithm to output the 3-4
set of all hypotheses consistent with the training examples.
Write a program to demonstrate the working of the decision tree
3 based ID3 algorithm. Use an appropriate data set for building the 5-8
decision tree and apply this knowledge toclassify a new sample
Build an Artificial Neural Network by implementing the
4 Backpropagation algorithm and test the same using appropriate 9-10
data sets
Write a program to implement the naïve Bayesian classifier for a
5 sample training data set. Compute the accuracy of the classifier, 11-13
considering few test data sets.
Assuming a set of documents that need to be classified, use the
6 naive Bayesian Classifier model to perform this task. Calculate 14-16
the accuracy, precision, and recall for your data set.
Write a program to construct a Bayesian network considering
7 medical data. Use this model to demonstrate the diagnosis of heart 17-18
patients using standard Heart Disease Data Set.
Apply EM algorithm to cluster a set of data set. Use the same data
8 set for clustering using k-Means algorithm. Compare the results 19-21
of these two algorithms.
Write a program to implement k-Nearest Neighbour algorithm to
9 22-23
classify the iris data set. Print both correct and wrong predictions.
Implement the non-parametric Locally Weighted Regression
10 algorithm in order to fit data points. Select appropriate data set 24-25
for your experiment and draw graphs.
Department of Computer Science & Engineering, MITRC, Alwar i

LAB INSTRUCTIONS
1. Keep silence and sit properly in the lab.
2. Keep your bag properly inside the lab.
3. Always follow the instruction given by concerned faulty to perform the assigned
experiment.
4. Do not turn on the PCs without permission.
5. Do not switch off the power supply of the PCs directly, first shut down the PCs then
switch off power supply.
6. Every student is responsible for any damage to the PCs or its accessories which is
assigned for lab work.
7. Installation or un-installation of any software is strictly prohibited.
8. Always bring your lab file and the task assigned to you must be completed.
9. Experiment performed by you should be positively checked in next turn after that
faculty may not check your work.
10. Please mention your roll number, name, node number and signature in lab register.
11. Do not go to assist any student.
Department of Computer Science & Engineering, MITRC, Alwar ii

BTU SYLLABUS
S.No Content
Implement and demonstrate the FIND-S algorithm for finding the most
1
specific hypothesis based on a given set of training data samples.
For a given set of training data examples, implement and demonstrate the
2 Candidate-Elimination algorithm to output a description of the set of all
hypotheses consistent with the training examples.
Write a program to demonstrate the working of the decision tree based ID3
3 algorithm. Use an appropriate data set for building the decision tree and apply
this knowledge to classify a new sample
Build an Artificial Neural Network by implementing the Backpropagation
4
algorithm and test the same using appropriate data sets
Write a program to implement the naive Bayesian classifier for a sample
5 training data set. Compute the accuracy of the classifier, considering few
test data sets.
Assuming a set of documents that need to be classified, use the naive Bayesian
Classifier model to perform this task. Built-in Java classes/API can be used to
6
write the program. Calculate the accuracy, precision, and recall for your data
set.
Write a program to construct aBayesian network considering medical data.
7 Use this model to demonstrate the diagnosis of heart patients using standard
Heart Disease Data Set. You can use Java/Python ML library classes/API.
Apply EM algorithm to cluster a set of data. Use the same data set
for clustering using k-Means algorithm. Compare the results of these
8
two algorithms and comment on the quality of clustering. You can add
Java/Python ML library classes/API in the program.
Write a program to implement k-Nearest Neighbour algorithm to classify the
9 iris data set. Print both correct and wrong predictions. Java/Python ML library
classes can be used for this problem.
Implement the non-parametric Locally Weighted Regression algorithm in
10
order to fit data points. Select appropriate data set and draw graphs.
Department of Computer Science & Engineering, MITRC, Alwar iii

LAB INTRODUCTION
This lab is intended for the third year students of engineering branches in the subject
of Machine Learning. This manual typically contains practical/Lab Sessions related ML
covering various aspects related to the subject to enhance understanding.
The programs are implemented in python programming language and involve use of
packages like numpy, pandas, matplotlib, scikit-learn.
Department of Computer Science & Engineering, MITRC, Alwar iv

Program No. 1
Objective: Implement and demonstrate the FIND-S algorithm for finding the most specific
hypothesis based on a given set of training data samples. Read the training data from a
.CSV file.
Program Python implementation of FIND-S algorithm

1 import csv
2 a = []
3
4 with open(’enjoysport.csv’, ’r’) as csvfile:
5 for row in csv.reader(csvfile):
6 a.append(row)
7 print(a)
8
9
10
11
12
num_attribute = len(a[0])-1
RC
print("\n The total number of training instances are : ",len(a))
13 print("\n The initial hypothesis is : ")

14 hypothesis = [’0’]*num_attribute
15 print(hypothesis)
IT
16
17 for i in range(0, len(a)):
18 if a[i][num_attribute] == ’yes’:
19 for j in range(0, num_attribute):
20 if hypothesis[j] == ’0’ or hypothesis[j] == a[i][j]:
21 hypothesis[j] = a[i][j]
M
22 else:
23 hypothesis[j] = ’?’
24 print("\n The hypothesis for the training instance {} is : \n".
format(i+1),hypothesis)
25
26 print("\n The Maximally specific hypothesis for the training instance
is ")
27 print(hypothesis)
Listing 1: Find-S program
Output
The Given Training Data Set
[’sunny’, ’warm’, ’normal’, ’strong’, ’warm’, ’same’, ’yes’]
[’sunny’, ’warm’, ’high’, ’strong’, ’warm’, ’same’, ’yes’]
[’rainy’, ’cold’, ’high’, ’strong’, ’warm’, ’change’, ’no’]
[’sunny’, ’warm’, ’high’, ’strong’, ’cool’, ’change’, ’yes’]
The total number of training instances are : 4
The initial hypothesis is :
Department of Computer Science & Engineering, MITRC, Alwar 1

[’0’, ’0’, ’0’, ’0’, ’0’, ’0’]

The hypothesis for the training instance 1 is :
[’sunny’, ’warm’, ’normal’, ’strong’, ’warm’, ’same’]
[’sunny’, ’warm’, ’?’, ’strong’, ’warm’, ’same’]
[’sunny’, ’warm’, ’?’, ’strong’, ’warm’, ’same’]
[’sunny’, ’warm’, ’?’, ’strong’, ’?’, ’?’]
The Maximally specific hypothesis for the training instance is
[’sunny’, ’warm’, ’?’, ’strong’, ’?’, ’?’]
Viva Voce
1. What do you mean by Concept learning?
RC
2. What do you mean by Inductive logic?
3. What do you understand by Hypothesis space?
4. Can Concept learning be used as searching algorithm?
5. Explain the Find-S Algorithm.

IT
Reference
1. Raschka, Sebastian, and Vahid Mirjalili. Python machine learning: Machine learning
and deep learning with Python, scikit-learn, and TensorFlow 2. Packt Publishing
M
Ltd, 2019.
2. Liu, Yuxi Hayden. Python Machine Learning By Example: Implement machine

learning algorithms and techniques to build intelligent systems. Packt Publishing
Ltd, 2019.

Program No. 2
Objective: For a given set of training data examples stored in a .CSV file, implement
and demonstrate the Candidate-Elimination algorithm to output a description of the set of
all hypotheses consistent with the training examples.
Program Python implementation of Candidate-Elimination algorithm

1 import numpy as np
2 import pandas as pd
3
4 data = pd.read_csv(’enjoysport.csv’)
5 concepts = np.array(data.iloc[:,0:-1])
6 print(concepts)
7 target = np.array(data.iloc[:,-1])
8 print(target) RC
9 def learn(concepts, target):
10 specific_h = concepts[0].copy()
11 print("initialization of specific_h and general_h")
12 print(specific_h)
13 general_h = [["?" for i in range(len(specific_h))] for i in range(
len(specific_h))]
14 print(general_h)
IT
15
16 for i, h in enumerate(concepts):
17 print("For Loop Starts")
18 if target[i] == "yes":
19 print("If instance is Positive ")
20 for x in range(len(specific_h)):
M
21 if h[x]!= specific_h[x]:
22 specific_h[x] =’?’
23 general_h[x][x] =’?’
24
25 if target[i] == "no":
26 print("If instance is Negative ")
27 for x in range(len(specific_h)):
28 if h[x]!= specific_h[x]:
29 general_h[x][x] = specific_h[x]
30 else:
31 general_h[x][x] = ’?’
32
33 print(" steps of Candidate Elimination Algorithm",i+1)

34 print(specific_h)
35 print(general_h)
36 print("\n")
37 print("\n")
38
39 indices = [i for i, val in enumerate(general_h) if val == [’?’, ’?’

, ’?’, ’?’, ’?’, ’?’]]
40 for i in indices:
41 general_h.remove([’?’, ’?’, ’?’, ’?’, ’?’, ’?’])
42 return specific_h, general_h

43
44 s_final, g_final = learn(concepts, target)

45
46 print("Final Specific_h:", s_final, sep="\n")
47 print("Final General_h:", g_final, sep="\n")
Listing 2: Candidate-Elimination program
Output
Final Specific_h:
[’sunny’ ’warm’ ’?’ ’strong’ ’?’ ’?’]
Final General_h:
[[’sunny’, ’?’, ’?’, ’?’, ’?’, ’?’],
[’?’, ’warm’, ’?’, ’?’, ’?’, ’?’]]
Viva Voce
RC
1. What do you mean by consistent hypothesis?
2. What do you mean by version space?
3. Explain Candidate-Elimination ALgorithm.?

IT
4. How Candidate-Elimination is different from Find-S algorithm.
Reference
M
Ltd, 2019.

Ltd, 2019.

Program No. 3
Objective: Write a program to demonstrate the working of the decision tree based ID3
algorithm. Use an appropriate data set for building the decision tree and apply this knowledge
to classify a new sample
Program Python implementation of decision tree based on ID3 algorithm

1 import math
2 import csv
3 def load_csv(filename):
4 lines=csv.reader(open(filename,"r"));
5 dataset = list(lines)
6 headers = dataset.pop(0)
7 return dataset,headers
8
9
10
11
12
class Node:
def __init__(self,attribute):
self.attribute=attribute
self.children=[]
RC
13 self.answer=""
14
15 def subtables(data,col,delete):
dic={}
IT
16
17 coldata=[row[col] for row in data]

18 attr=list(set(coldata))
19
20 counts=[0]*len(attr)
21 r=len(data)
M
22 c=len(data[0])
23 for x in range(len(attr)):
24 for y in range(r):
25 if data[y][col]==attr[x]:
26 counts[x]+=1
27
29 dic[attr[x]]=[[0 for i in range(c)] for j in range(counts[x])]
30 pos=0
31 for y in range(r):
32 if data[y][col]==attr[x]:
33 if delete:
34 del data[y][col]
35 dic[attr[x]][pos]=data[y]
36 pos+=1
37 return attr,dic
38
39 def entropy(S):
40 attr=list(set(S))
41 if len(attr)==1:
42 return 0
43
44 counts=[0,0]

45 for i in range(2):
46 counts[i]=sum([1 for x in S if attr[i]==x])/(len(S)*1.0)
47
48 sums=0
49 for cnt in counts:
50 sums+=-1*cnt*math.log(cnt,2)
51 return sums
52
53 def compute_gain(data,col):
54 attr,dic = subtables(data,col,delete=False)
55
56 total_size=len(data)
57 entropies=[0]*len(attr)
58 ratio=[0]*len(attr)
59
60 total_entropy=entropy([row[-1] for row in data])
62 ratio[x]=len(dic[attr[x]])/(total_size*1.0)
63 entropies[x]=entropy([row[-1] for row in dic[attr[x]]])
64
65
66
67
return total_entropy
RC
total_entropy-=ratio[x]*entropies[x]
def build_tree(data,features):
68 lastcol=[row[-1] for row in data]
69 if(len(set(lastcol)))==1:
70 node=Node("")
71 node.answer=lastcol[0]
IT
72 return node
73
74 n=len(data[0])-1
75 gains=[0]*n
76 for col in range(n):
M
77 gains[col]=compute_gain(data,col)
78 split=gains.index(max(gains))
79 node=Node(features[split])
80 fea = features[:split]+features[split+1:]
81
82
83 attr,dic=subtables(data,split,delete=True)
84
86 child=build_tree(dic[attr[x]],fea)
87 node.children.append((attr[x],child))
88 return node
89
90 def print_tree(node,level):
91 if node.answer!="":
92 print(" "*level,node.answer)
93 return
94
95 print(" "*level,node.attribute)
96 for value,n in node.children:
97 print(" "*(level+1),value)
98 print_tree(n,level+2)
99

100
101 def classify(node,x_test,features):

102 if node.answer!="":
103 print(node.answer)
104 return
105 pos=features.index(node.attribute)
106 for value, n in node.children:
107 if x_test[pos]==value:
108 classify(n,x_test,features)
109
110 ’’’Main program’’’
111 dataset,features=load_csv("id3.csv")
112 node1=build_tree(dataset,features)
113
114 print("The decision tree for the dataset using ID3 algorithm is")
115 print_tree(node1,0)
116 testdata,features=load_csv("id3_test.csv")
117
118 for xtest in testdata:

119
120
121
RC
print("The test instance:",xtest)
print("The label for test instance:",end="
classify(node1,xtest,features)
")
Listing 3: Decision tree using ID3 algorithm
Output
IT
The decision tree for the dataset using ID3 algorithm is
Outlook
rain
M
Wind
strong
no
weak
yes
overcast
yes
sunny
Humidity
normal
yes
high
no
The test instance: [’rain’, ’cool’, ’normal’, ’strong’]
The label for test instance: no
Viva Voce
1. How you decide the root node of decision tree?

2. What is role of entropy in decision tree?
3. what is pruning of decision tree?
Reference
Ltd, 2019.

Ltd, 2019.
RC
IT
M

Program No. 4
Objective: Build an Artificial Neural Network by implementing the Backpropagation

algorithm and test the same using appropriate data sets
Program ANN using backpropagation algorithm implemented in python language

2 X = np.array(([2, 9], [1, 5], [3, 6]), dtype=float) # two inputs [sleep
,study]
3 y = np.array(([92], [86], [89]), dtype=float) # one output [Expected %
in Exams]
4 X = X/np.amax(X,axis=0) # maximum of X array longitudinally
5 y = y/100
6
7 #Sigmoid Function RC
8 def sigmoid (x):
9 return 1/(1 + np.exp(-x))
10
11 #Derivative of Sigmoid Function
12 def derivatives_sigmoid(x):
13 return x * (1 - x)
14
#Variable initialization
IT
15
16 epoch=5000 #Setting training iterations
17 lr=0.1 #Setting learning rate
18 inputlayer_neurons = 2 #number of features in data set
19 hiddenlayer_neurons = 3 #number of hidden layers neurons
20 output_neurons = 1 #number of neurons at output layer
M
21
22 #weight and bias initialization
23 wh=np.random.uniform(size=(inputlayer_neurons,hiddenlayer_neurons)) #
weight of the link from input node to hidden node
24 bh=np.random.uniform(size=(1,hiddenlayer_neurons)) # bias of the link
from input node to hidden node
25 wout=np.random.uniform(size=(hiddenlayer_neurons,output_neurons)) #
weight of the link from hidden node to output node
26 bout=np.random.uniform(size=(1,output_neurons)) #bias of the link from
hidden node to output node
27
28
29 #draws a random range of numbers uniformly of dim x*y
30 for i in range(epoch):
31
32 #Forward Propogation
33 hinp1=np.dot(X,wh)
34 hinp=hinp1 + bh
35 hlayer_act = sigmoid(hinp)
36 outinp1=np.dot(hlayer_act,wout)
37 outinp= outinp1+ bout
38 output = sigmoid(outinp)
39

40 #Backpropagation
41 EO = y-output
42 outgrad = derivatives_sigmoid(output)
43 d_output = EO* outgrad
44 EH = d_output.dot(wout.T)
45
46 #how much hidden layer weights contributed to error
47 hiddengrad = derivatives_sigmoid(hlayer_act)
48 d_hiddenlayer = EH * hiddengrad
49
50 # dotproduct of nextlayererror and currentlayerop
51 wout += hlayer_act.T.dot(d_output) *lr
52 wh += X.T.dot(d_hiddenlayer) *lr
53
54 print("Input: \n" + str(X))
55 print("Actual Output: \n" + str(y))
56 print("Predicted Output: \n" ,output)
Listing 4: ANN backpropogation algorithm implementation
Output
RC
Viva Voce
1. What is pereceptron?
IT
2. how multi-layer ANN is better than single layer ANN?
3. what is Back-Propogation Algorithm?

M
Reference
Ltd, 2019.

Ltd, 2019.

Program No. 5
Objective: Write a program to implement the naive Bayesian classifier for a sample
training data set stored as a .CSV file. Compute the accuracy of the classifier, considering
few test data sets
Program Naive Bayesian classifier implemented in python

1 import csv
2 import random
3 import math
4
5 def loadcsv(filename):
6 lines = csv.reader(open(filename, "r"));
7 dataset = list(lines)
8 for i in range(len(dataset)): RC
9 #converting strings into numbers for processing
10 dataset[i] = [float(x) for x in dataset[i]]
11
12 return dataset
13
14 def splitdataset(dataset, splitratio):
15 #67% training size
trainsize = int(len(dataset) * splitratio);
IT
16
17 trainset = []
18 copy = list(dataset);
19 while len(trainset) < trainsize:
20 #generate indices for the dataset list randomly to pick ele for
training data
M
21 index = random.randrange(len(copy));
22 trainset.append(copy.pop(index))
23 return [trainset, copy]
24
25 def separatebyclass(dataset):
26 separated = {} #dictionary of classes 1 and 0
27 #creates a dictionary of classes 1 and 0 where the values are
28 #the instances belonging to each class
29 for i in range(len(dataset)):
30 vector = dataset[i]
31 if (vector[-1] not in separated):
32 separated[vector[-1]] = []
33 separated[vector[-1]].append(vector)
34 return separated
35
36 def mean(numbers):
37 return sum(numbers)/float(len(numbers))
38
39 def stdev(numbers):
40 avg = mean(numbers)
41 variance = sum([pow(x-avg,2) for x in numbers])/float(len(numbers)-1)
42 return math.sqrt(variance)
43

44 def summarize(dataset): #creates a dictionary of classes

45 summaries = [(mean(attribute), stdev(attribute)) for attribute in zip
(*dataset)];
46 del summaries[-1] #excluding labels +ve or -ve
47 return summaries
48
49 def summarizebyclass(dataset):
50 separated = separatebyclass(dataset);
51 #print(separated)
52 summaries = {}
53 for classvalue, instances in separated.items():
54 #for key,value in dic.items()
55 #summaries is a dic of tuples(mean,std) for each class value
56 summaries[classvalue] = summarize(instances) #summarize is used to
cal to mean and std
57 return summaries
58
59 def calculateprobability(x, mean, stdev):
60 exponent = math.exp(-(math.pow(x-mean,2)/(2*math.pow(stdev,2))))
61
62
63
64
RC
return (1 / (math.sqrt(2*math.pi) * stdev)) * exponent
def calculateclassprobabilities(summaries, inputvector):

probabilities = {} # probabilities contains the all prob of all class
of test data
65 for classvalue, classsummaries in summaries.items():#class and
attribute information as mean and sd
66 probabilities[classvalue] = 1
IT
67 for i in range(len(classsummaries)):
68 mean, stdev = classsummaries[i] #take mean and sd of every
attribute for class 0 and 1 seperaely
69 x = inputvector[i] #testvector’s first attribute
70 probabilities[classvalue] *= calculateprobability(x, mean, stdev)
M
;#use normal dist

71 return probabilities
72
73 def predict(summaries, inputvector): #training and test data is passed

74 probabilities = calculateclassprobabilities(summaries, inputvector)
75 bestLabel, bestProb = None, -1
76 for classvalue, probability in probabilities.items():#assigns that
class which has he highest prob
77 if bestLabel is None or probability > bestProb:
78 bestProb = probability
79 bestLabel = classvalue
80 return bestLabel
81
82 def getpredictions(summaries, testset):
83 predictions = []
84 for i in range(len(testset)):
85 result = predict(summaries, testset[i])
86 predictions.append(result)
87 return predictions
88
89 def getaccuracy(testset, predictions):
90 correct = 0
91 for i in range(len(testset)):

92 if testset[i][-1] == predictions[i]:
93 correct += 1
94 return (correct/float(len(testset))) * 100.0
95
96 def main():
97 filename = ’naivedata.csv’
98 splitratio = 0.67
99 dataset = loadcsv(filename);
100
101 trainingset, testset = splitdataset(dataset, splitratio)
102 print(’Split {0} rows into train={1} and test={2} rows’.format(len(
dataset), len(trainingset), len(testset)))
103 # prepare model
104 summaries = summarizebyclass(trainingset);
105 #print(summaries)
106 # test model
107 predictions = getpredictions(summaries, testset) #find the
predictions of test data with the training data
108 accuracy = getaccuracy(testset, predictions)
109
110
111 main()
RC
print(’Accuracy of the classifier is : {0}%’.format(accuracy))
Listing 5: Bayesian Classifier implementation
Output
IT
Split 768 rows into train=514 and test=254 rows
Accuracy of the classifier is : 71.65354330708661%
M
Viva Voce
1. What do you understand by classification Algorithm in machine learning?
2. What do you mean by conditional probability?
3. Definte Bayes Theorem?
4. Explain how Bayes theorem is used for classification?
Reference
Ltd, 2019.

Ltd, 2019.

Program No. 6
Objective: Assuming a set of documents that need to be classified, use the naive Bayesian
Classifier model to perform this task.Calculate the accuracy, precision, and recall for your
data set.
Program Naive Bayesian classifier implemented using python Scikit-learn package.

2 from sklearn.model_selection import train_test_split
3 from sklearn.feature_extraction.text import CountVectorizer
4 from sklearn.naive_bayes import MultinomialNB
5 from sklearn import metrics
6
7 msg=pd.read_csv(’naivetext.csv’,names=[’message’,’label’])
8
9
10
11
12
RC
print(’The dimensions of the dataset’,msg.shape)
msg[’labelnum’]=msg.label.map({’pos’:1,’neg’:0})
X=msg.message
13 y=msg.labelnum
14
15 #splitting the dataset into train and test data
IT
16 xtrain,xtest,ytrain,ytest=train_test_split(X,y)
17 print (’\n the total number of Training Data :’,ytrain.shape)
18 print (’\n the total number of Test Data :’,ytest.shape)
19
20
21 #output the words or Tokens in the text documents
M
22 cv = CountVectorizer()
23 xtrain_dtm = cv.fit_transform(xtrain)
24 xtest_dtm=cv.transform(xtest)
25 print(’\n The words or Tokens in the text documents \n’)
26 print(cv.get_feature_names())
27 df=pd.DataFrame(xtrain_dtm.toarray(),columns=cv.get_feature_names())
28
29 # Training Naive Bayes (NB) classifier on training data.
30 clf = MultinomialNB().fit(xtrain_dtm,ytrain)
31 predicted = clf.predict(xtest_dtm)
32
33 #printing accuracy, Confusion matrix, Precision and Recall

34 print(’\n Accuracy of the classifier is’,metrics.accuracy_score(ytest,
predicted))
35 print(’\n Confusion matrix’)
36 print(metrics.confusion_matrix(ytest,predicted))
37 print(’\n The value of Precision’, metrics.precision_score(ytest,
predicted))
38 print(’\n The value of Recall’, metrics.recall_score(ytest,predicted))
Listing 6: Bayesian Classifier

Output
The dimensions of the dataset (18, 2)
0 I love this sandwich

1 This is an amazing place
2 I feel very good about these beers
3 This is my best work
4 What an awesome view
5 I do not like this restaurant
6 I am tired of this stuff
7 I can’t deal with this
8 He is my sworn enemy
9 My boss is horrible
10 This is an awesome place
11 I do not like the taste of this juice
12
13
14
15
I love to dance RC
I am sick and tired of this place
What a great holiday
That is a bad locality to stay
16 We will have good fun tomorrow
17 I went to my enemy’s house today
IT
Name: message, dtype: object
0 1
1 1
M
2 1
3 1
4 1
5 0
6 0
7 0
8 0
9 0
10 1
11 0
12 1
13 0
14 1
15 0
16 1
17 0
Name: labelnum, dtype: int64

The total number of Training Data: (13,)

The total number of Test Data: (5,)
The words or Tokens in the text documents
[’about’, ’am’, ’amazing’, ’an’, ’and’, ’awesome’, ’beers’,

’best’, ’can’, ’deal’, ’do’, ’enemy’, ’feel’,
’fun’, ’good’, ’great’, ’have’, ’he’, ’holiday’, ’house’,
’is’, ’like’, ’love’, ’my’, ’not’, ’of’, ’place’,
’restaurant’, ’sandwich’, ’sick’, ’sworn’, ’these’, ’this’,
’tired’, ’to’, ’today’, ’tomorrow’, ’very’,
’view’, ’we’, ’went’, ’what’, ’will’, ’with’, ’work’]
Accuracy of the classifier is 0.8
Confusion matrix RC
[[2 1]
[0 2]]
The value of Precision 0.6666666666666666

The value of Recall 1.0
IT
Viva Voce
1. What do you understand by classification Algorithm in machine learning?
2. What do you mean by conditional probability?

M
3. Definte Bayes Theorem?
4. Explain how Bayes theorem is used for classification?
Reference
Ltd, 2019.

Ltd, 2019.

Program No. 7
Objective: Write a program to construct a Bayesian network considering medical data.

Use this model to demonstrate the diagnosis of heart patients using standard Heart Disease
Data Set.
Program Bayesian network constructed using python pgmpy package used for making
network.
2 import csv
4 from pgmpy.models import BayesianModel
5 from pgmpy.estimators import MaximumLikelihoodEstimator
6 from pgmpy.inference import VariableElimination
7
8
9
10
11
RC
heartDisease = pd.read_csv(’heart.csv’)
heartDisease = heartDisease.replace(’?’,np.nan)
print(’Few examples from the dataset are given below’)

12 print(heartDisease.head())
13
14 model = BayesianModel([(’age’,’trestbps’),(’age’,’fbs’),(’sex’,’
IT
trestbps’),
15 (’exang’,’trestbps’),(’trestbps’,’heartdisease’)
,
16 (’fbs’,’heartdisease’),(’heartdisease’,’restecg’
),
17 (’heartdisease’,’thalach’),(’heartdisease’,’chol
M
’)])
18
19 print(’\nLearning CPD using Maximum likelihood estimators’)

20 model.fit(heartDisease,estimator=MaximumLikelihoodEstimator)
21
22 print(’\n Inferencing with Bayesian Network:’)
23 HeartDisease_infer = VariableElimination(model)
24
25 print(’\n 1. Probability of HeartDisease given Age=28’)

26 q=HeartDisease_infer.query(variables=[’heartdisease’],evidence={’age’
:28})
27 print(q[’heartdisease’])
28
29 print(’\n 2. Probability of HeartDisease given cholesterol=100’)

30 q=HeartDisease_infer.query(variables=[’heartdisease’],evidence={’chol’
:100})
31 print(q[’heartdisease’])
Listing 7: Bayesian Network

Output
RC
IT
Viva Voce
M
1. How you construct Bayes Network?
2. What are the advantages of Bayes Network ?
Reference
Ltd, 2019.

Ltd, 2019.

Program No. 8
Objective: Apply EM algorithm to cluster a set of data stored in a .CSV file. Use the
same data set for clustering using k-Means algorithm. Compare the results of these two
algorithms and comment on the quality of clustering.
Program K-means algorithm implemented using python scikit-learn package.

1 import matplotlib.pyplot as plt
2 from sklearn import datasets
3 from sklearn.cluster import KMeans
4 import sklearn.metrics as sm
7 #import matplotlib inline
8
9
10
11
12
iris = datasets.load_iris()
X = pd.DataFrame(iris.data)
RC
X.columns = [’Sepal_Length’,’Sepal_Width’,’Petal_Length’,’Petal_Width’]
13
14 y = pd.DataFrame(iris.target)
15 y.columns = [’Targets’]
IT
16
17 #colormap = np.array([’red’, ’lime’, ’black’])

18
19 # K Means Cluster
20 model = KMeans(n_clusters=3)
21 model.fit(X)
M
22 # This is what KMeans thought

23 model.labels_
24
25 # View the results
26
27 # Set the size of the plot
28 plt.figure(figsize=(14,7))
29
30 # Create a colormap
31 colormap = np.array([’red’, ’lime’, ’black’])
32
33 # Plot the Original Classifications
34 plt.subplot(1, 2, 1)
35 plt.scatter(X.Petal_Length, X.Petal_Width, c=colormap[y.Targets], s=40)
36 plt.title(’Real Classification’)
37
38 # Plot the Models Classifications
40 plt.scatter(X.Petal_Length, X.Petal_Width, c=colormap[model.labels_], s
=40)
41 plt.title(’K Mean Classification’)
42
43 # View the results

44 # Set the size of the plot

45 plt.figure(figsize=(14,7))
46 # Create a colormap
47 #print(’The accuracy score : ’,sm.accuracy_score(y, model.labels_))
48 #sm.confusion_matrix(y, model.labels_)
49
50 predY = np.choose(model.labels_, [0, 1, 2]).astype(np.int64)
51 print (predY)
52
53 #colormap = np.array([’red’, ’lime’, ’black’])
54 # Plot Orginal
56 plt.scatter(X.Petal_Length, X.Petal_Width, c=colormap[y.Targets], s=40)
57 plt.title(’Real Classification’)
58 # Plot Predicted with corrected values
60 plt.scatter(X.Petal_Length,X.Petal_Width, c=colormap[predY], s=40)
61 plt.title(’K Mean Classification’)
62
63
64
labels_))
RC
print(’The accuracy score of K-Mean: ’,sm.accuracy_score(y, model.
print(’The Confusion matrixof K-Mean: ’,sm.confusion_matrix(y, model.

labels_))
65
66 from sklearn import preprocessing

67 scaler = preprocessing.StandardScaler()
68 scaler.fit(X)
IT
69 xsa = scaler.transform(X)
70 xs = pd.DataFrame(xsa, columns = X.columns)
71 #xs.sample(5)
72
73 from sklearn.mixture import GaussianMixture
M
74 gmm = GaussianMixture(n_components=3)
75 gmm.fit(xs)
76
77 y_cluster_gmm = gmm.predict(xs)
78 #y_cluster_gmm
79
81 plt.scatter(X.Petal_Length, X.Petal_Width, c=colormap[y_cluster_gmm], s
=40)
82 plt.title(’GMM Classification’)
83
84 print(’The accuracy score of EM: ’,sm.accuracy_score(y, y_cluster_gmm))
85 print(’The Confusion matrix of EM: ’,sm.confusion_matrix(y,
y_cluster_gmm))
Listing 8: Clustering
Viva Voce
1. Is clustering unsupervised or supervised learning ?
2. Name few important clustering algorithm.

3. What are advantages of k-means clustering?
Reference
Ltd, 2019.

Ltd, 2019.
RC
IT
M

Program No. 9
Objective: Write a program to implement k-Nearest Neighbour algorithm to classify the

iris data set. Print both correct and wrong predictions.
Program KNN algorithm implemented using python scikit-learn package.

1 from sklearn.model_selection import train_test_split
2 from sklearn.neighbors import KNeighborsClassifier
3 from sklearn.metrics import classification_report, confusion_matrix
4 from sklearn import datasets
5
6 iris=datasets.load_iris()
7
8 x = iris.data
9 y = iris.target RC
10
11 print (’sepal-length’, ’sepal-width’, ’petal-length’, ’petal-width’)
12 print(x)
13 print(’class: 0-Iris-Setosa, 1- Iris-Versicolour, 2- Iris-Virginica’)
14 print(y)
15
16 x_train, x_test, y_train, y_test = train_test_split(x,y,test_size=0.3)
IT
17
18 #To Training the model and Nearest nighbors K=5
19 classifier = KNeighborsClassifier(n_neighbors=5)
20 classifier.fit(x_train, y_train)
21
22 #To make predictions on our test data
M
23 y_pred=classifier.predict(x_test)
24
25 print(’Confusion Matrix’)
26 print(confusion_matrix(y_test,y_pred))
27 print(’Accuracy Metrics’)
28 print(classification_report(y_test,y_pred))
Listing 9: KNN implemenation
Output
sepal-length sepal-width petal-length petal-width
[[5.1 3.5 1.4 0.2]
[4.9 3. 1.4 0.2]
[4.7 3.2 1.3 0.2]
[4.6 3.1 1.5 0.2]
[5. 3.6 1.4 0.2]
. . . . .
. . . . .
[6.2 3.4 5.4 2.3]

[5.9 3. 5.1 1.8]]

class: 0-Iris-Setosa, 1- Iris-Versicolour, 2- Iris-Virginica
[0 0 0 ......0 0 1 1 1 ........1 1 2 2 2 ........ 2 2]
Confusion Matrix[[20 0 0]
[ 0 10 0]
[ 0 1 14]]
Accuracy Metrics
Precision recall f1-score support
0 1.00 1.00 1.00 20
1 0.91 1.00 0.95 10
2 1.00 0.93 0.97 15
avg / total 0.98 0.98 0.98 45
Viva Voce RC
1. What are features of iris-data?
2. Explain K-nearest neighbour algorithm?
3. K-nearest neighbour is regression or classification?

IT
Reference
M
Ltd, 2019.

Ltd, 2019.

Program No. 10
Objective: Implement the non-parametric Locally Weighted Regression algorithm in

order to fit data points. Select appropriate data set for your experiment and draw graphs.
Program Non-parametric regression implemented using python language.

2 from bokeh.plotting import figure, show, output_notebook
3 from bokeh.layouts import gridplot
4 from bokeh.io import push_notebook
5
6 def local_regression(x0, X, Y, tau):# add bias term
7 x0 = np.r_[1, x0] # Add one to avoid the loss in information
8 X = np.c_[np.ones(len(X)), X]
9 # fit model: normal equations with kernel
RC
10 xw = X.T * radial_kernel(x0, X, tau) # XTranspose * W
11 beta = np.linalg.pinv(xw * X) * xw * Y #@ Matrix Multiplication or
Dot Product
12 # predict value
13 return x0 * beta # @ Matrix Multiplication or Dot Product for
prediction
14
IT
15
16 def radial_kernel(x0, X, tau):
17 return np.exp(np.sum((X - x0) ** 2, axis=1) / (-2 * tau * tau))
18 # Weight or Radial Kernal Bias Function
19
20 n = 1000
M
21 # generate dataset
22 X = np.linspace(-3, 3, num=n)
23 print("The Data Set ( 10 Samples) X :\n",X[1:10])
24 Y = np.log(np.abs(X ** 2 - 1) + .5)
25 print("The Fitting Curve Data Set (10 Samples) Y :\n",Y[1:10])
26 # jitter X
27 X += np.random.normal(scale=.1, size=n)
28 print("Normalised (10 Samples) X :\n",X[1:10])
29
30 domain = np.linspace(-3, 3, num=300)

31 print(" Xo Domain Space(10 Samples) :\n",domain[1:10])
32
33 def plot_lwr(tau):
34 # prediction through regression
35 prediction = [local_regression(x0, X, Y, tau) for x0 in domain]
36 plot = figure(plot_width=400, plot_height=400)
37 plot.title.text=’tau=%g’ % tau
38 plot.scatter(X, Y, alpha=.3)
39 plot.line(domain, prediction, line_width=2, color=’red’)
40 return plot
41
42 show(gridplot([
43 [plot_lwr(10.), plot_lwr(1.)],

44 [plot_lwr(0.1), plot_lwr(0.01)]]))
Listing 10: Non-parametric regression
Output
RC
IT
M
Viva Voce
1. What is difference between regression and classification?
2. Is regression supervised learning?
3. Name few important regression algorithm.
Reference
Ltd, 2019.
Ltd, 2019.

Lab Manual: Department of Computer Science and Engineering

Uploaded by

Copyright:

Available Formats

Lab Manual: Department of Computer Science and Engineering

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lab Manual: Department of Computer Science and Engineering

Uploaded by

Copyright:

Available Formats

DEPARTMENT OF COMPUTER SCIENCE

MACHINE LEARNING LAB

Modern Institute of Technology & Research Centre,

Department of Computer Science & Engineering, MITRC, Alwar i

1. Keep silence and sit properly in the lab.

2. Keep your bag properly inside the lab.

4. Do not turn on the PCs without permission.

7. Installation or un-installation of any software is strictly prohibited.

11. Do not go to assist any student.

Department of Computer Science & Engineering, MITRC, Alwar ii

Department of Computer Science & Engineering, MITRC, Alwar iii

Department of Computer Science & Engineering, MITRC, Alwar iv

Program Python implementation of FIND-S algorithm

13 print("\n The initial hypothesis is : ")

Department of Computer Science & Engineering, MITRC, Alwar 1

[’0’, ’0’, ’0’, ’0’, ’0’, ’0’]

3. What do you understand by Hypothesis space?

4. Can Concept learning be used as searching algorithm?

5. Explain the Find-S Algorithm.

2. Liu, Yuxi Hayden. Python Machine Learning By Example: Implement machine

Department of Computer Science & Engineering, MITRC, Alwar 2

Program Python implementation of Candidate-Elimination algorithm

33 print(" steps of Candidate Elimination Algorithm",i+1)

39 indices = [i for i, val in enumerate(general_h) if val == [’?’, ’?’

Department of Computer Science & Engineering, MITRC, Alwar 3

44 s_final, g_final = learn(concepts, target)

2. What do you mean by version space?

3. Explain Candidate-Elimination ALgorithm.?

2. Liu, Yuxi Hayden. Python Machine Learning By Example: Implement machine

Department of Computer Science & Engineering, MITRC, Alwar 4

Program Python implementation of decision tree based on ID3 algorithm

17 coldata=[row[col] for row in data]

Department of Computer Science & Engineering, MITRC, Alwar 5

Department of Computer Science & Engineering, MITRC, Alwar 6

101 def classify(node,x_test,features):

118 for xtest in testdata:

Listing 3: Decision tree using ID3 algorithm

Department of Computer Science & Engineering, MITRC, Alwar 7

2. What is role of entropy in decision tree?

3. what is pruning of decision tree?

2. Liu, Yuxi Hayden. Python Machine Learning By Example: Implement machine

Department of Computer Science & Engineering, MITRC, Alwar 8

Objective: Build an Artificial Neural Network by implementing the Backpropagation

Program ANN using backpropagation algorithm implemented in python language

Department of Computer Science & Engineering, MITRC, Alwar 9

3. what is Back-Propogation Algorithm?

2. Liu, Yuxi Hayden. Python Machine Learning By Example: Implement machine

Department of Computer Science & Engineering, MITRC, Alwar 10

Program Naive Bayesian classifier implemented in python

Department of Computer Science & Engineering, MITRC, Alwar 11

44 def summarize(dataset): #creates a dictionary of classes

def calculateclassprobabilities(summaries, inputvector):

;#use normal dist

73 def predict(summaries, inputvector): #training and test data is passed

Department of Computer Science & Engineering, MITRC, Alwar 12

Listing 5: Bayesian Classifier implementation

2. What do you mean by conditional probability?

3. Definte Bayes Theorem?