TM

MACHINE LEARNING
LABORATORY MANUAL
15CSL76
Department of Information Science & Engineering
Sahyadri College of Engineering & Management
Mangaluru

Authors : Praahas Amin & Madhura N. Hegde

Department of Information Science & Engineering
Sahyadri College of Engineering & Management
Mangaluru

‘
 1

PREFACE
This Machine Learning Laboratory Manual is designed for Information Science & Engineering
students, Sahyadri College of Engineering and Management, Mangaluru.

Machine Learning Algorithms have gained wide application in many domains. Therefore, this
laboratory manual is prepared with the intention of providing a learning opportunity for the
students to understand the basics of programming Machine Learning Algorithms in Python

The lab experiments enhance the learning curve for the students and give them hands-on
experience on the different concepts that were discussed in the theory classes.

The manual contains the prescribed experiments for easy & quick understanding of the
students. The authors have gathered materials from Books, Journals and Web resources.

We hope that this practical manual will be helpful for the students of Information Science &
Engineering for understanding the subject from the point of view of applied aspects.

There is always scope for improvement in the manual. We would appreciate to receive valuable
suggestions from readers and users for future use.

This Laboratory manual is based on the syllabus provided by VTU under the subject code
15CSL76.
 2

ACKNOWLEDGMENTS
We are highly indebted to Dr. Shamanth Rai. Professor & H.O.D. Department of Information
Science & Engineering for providing us the opportunity to prepare this manual as well as for
his guidance and constant supervision.

We would like to express our special gratitude and thanks to Dr. Srinivas Rao Kunte,
Principal Sahyadri College of Engineering & Management, Mangaluru.

We are also thankful to Dr. D. L. Prabhakara, Director Sahyadri Group of Institutions, and
Dr. Manjunath Bhandary, Chairman, Sahyadri Group of Institutions.

We are also thankful to Mr. Praveen Damle for proofreading and constant support in the
Laboratories.

We are grateful to Mr. Gautham G. Rao (4SF15IS028) for all the support provided in
development of the code.

We are thankful to the students and colleagues for their support and encouragement.

Praahas Amin
Madhura N. Hegde
 3

S.No. Experiments Page

Implement and demonstrate the FIND-S algorithm for finding the most specific
1 hypothesis based on a given set of training data samples. Read the training data 4
from a .CSV file.
For a given set of training data examples stored in a .CSV file, implement and
2 demonstrate the Candidate-Elimination algorithm to output a description of the 6
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 8
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. 11

Write a program to implement the Naïve Bayes classifier for a sample training
5 data set stored as a .CSV file. Compute the accuracy of the classifier, 13
considering few test data sets.
Assuming a set of documents that need to be classified, use the naïve Bayesian
6 Classifier model to perform this task. Built-in Java classes/API can be used to 16
write the program. Calculate the accuracy, precision, and recall for your data set.
Write a program to construct a Bayesian network considering medical data. Use
7 this model to demonstrate the diagnosis of heart patients using standard Heart 19
Disease Data Set. You can use Java/Python ML library classes/API.
Apply EM algorithm to cluster a set of data stored in a .CSV file. Use the same
8 data set for clustering using k-Means algorithm. Compare the results of these two 23
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 iris
9 data set. Print both correct and wrong predictions. Java/Python ML library 26
classes can be used for this problem.
Implement the non-parametric Locally Weighted Regression algorithm in order
10 to fit data points. Select appropriate data set for your experiment and draw 28
graphs.
All laboratory experiments are to be included for practical examination.
• Students are allowed to pick one experiment from the lot.
• Strictly follow the instructions as printed on the cover page of answer script
• Marks distribution: Procedure + Conduction + Viva:20 + 50 +10 (80)
Change of experiment is allowed only once and marks allotted to the procedure part to be made 0.
Subject Code: 15CSL76 | IA Marks: 20 | Exam Marks: 80 | Exam Hours: 03 | Credits: 02
 4

Experiment 1: Find-S Algorithm

Aim: 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.
import pandas as pd
from pandas import DataFrame
data = DataFrame.from_csv('EnjoySport.csv')
columnLength= data.shape[1]
print (data.values)
h = ['0']*(columnLength-1)
hp=[]
hn=[]
for trainingExample in data.values:
if trainingExample[-1]!='no':
hp.append(list(trainingExample))
else:
hn.append(list(trainingExample))
for i in range (len(hp)):
for j in range(columnLength-1):
if (h[j]=='0'):
h[j]=hp[i][j]
if (h[j]!=hp[i][j]):
h[j]='?'
else:
h[j]=hp[i][j]
print('\nThe positive Hypotheses are:',hp)
print('\nThe negative Hypotheses are:',hn)
print('\nThe Maximally Specific Hypothesis h is:',h)

Output:
 5

Dataset:

Source: https://github.com/praahas/machine-learning-vtu
 6

Experiment 2: Candidate-Elimination Algorithm

Aim: 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.
from pandas import DataFrame
data=DataFrame.from_csv('EnjoySport.csv')
concepts=data.values[:,:-1]
target=data.values[:,-1]

def learn(concepts, target):

specific_h = concepts[0].copy()
general_h = [['?' for i in range(len(specific_h))] for i in range(len(specific_h))]
for i, h in enumerate(concepts):
if target[i] == "yes":
#print(target[i])
for x in range(len(specific_h)):
if h[x] != specific_h[x]:
specific_h[x] = '?'
general_h[x][x] = '?'
if target[i] == "no":
for x in range(len(specific_h)):
if h[x] != specific_h[x]:
general_h[x][x] = specific_h[x]
else:
general_h[x][x] = '?'
indices = [i for i,val in enumerate(general_h) if val==['?' for i in range(len(specific_h))]]
for i in indices:
general_h.remove(['?' for i in range(len(specific_h))])
return specific_h, general_h

s_final, g_final = learn(concepts, target)

print("Final S:", s_final)
print("Final G:", g_final)

Output:
 7

Dataset:

Source: https://github.com/ggrao1/Candidate-Elimination
 8

Experiment 3: Decision Tree based ID3 Algorithm

Aim: 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.
def infoGain(P, N):
import math
return -P / (P + N) * math.log2(P / ( P + N)) - N / (P + N) * math.log2(N / (P + N))

def insertNode(tree, addTo, Node):

for k, v in tree.items():
if isinstance(v, dict):
tree[k] = insertNode(v, addTo, Node)
if addTo in tree:
if isinstance(tree[addTo], dict):
tree[addTo][Node] = 'None'
else:
tree[addTo] = {Node:'None'}
return tree

def insertConcept(tree, addTo, Node):

for k, v in tree.items():
if isinstance(v, dict):
tree[k] = insertConcept(v, addTo, Node)
if addTo in tree:
tree[addTo] = Node
return tree

def getNextNode(data, AttributeList, concept, conceptVals, tree, addTo):

Total = data.shape[0]
if Total == 0:
return tree

countC = {}
for cVal in conceptVals:
dataCC = data[data[concept] = = cVal]
countC[cVal] = dataCC.shape[0]

if countC[conceptVals[0]] = = 0:
tree = insertConcept(tree, addTo, conceptVals[1])
return tree

if countC[conceptVals[1]] = = 0:
tree = insertConcept(tree, addTo, conceptVals[0])
return tree

ClassEntropy = infoGain(countC[conceptVals[1]],countC[conceptVals[0]])
Attr = {}
for a in AttributeList:
Attr[a] = list(set(data[a]))
AttrCount = {}
EntropyAttr = {}
 9

for att in Attr:

for vals in Attr [att]:
for c in conceptVals:
iData = data[data[att] = = vals]
dataAtt = iData[iData[concept] = = c]
AttrCount[c] = dataAtt.shape[0]
TotalInfo = AttrCount[conceptVals[1]] + AttrCount[conceptVals[0]]
if AttrCount[conceptVals[1]] = = 0 or AttrCount[conceptVals[0]] = = 0:
InfoGain=0
else:
InfoGain = infoGain(AttrCount[conceptVals[1]], AttrCount[conceptVals[0]])

if att not in EntropyAttr:

EntropyAttr[att] = ( TotalInfo / Total ) * InfoGain
else:
EntropyAttr[att] = EntropyAttr[att] + ( TotalInfo / Total ) * InfoGain

Gain = {}
for g in EntropyAttr:
Gain[g] = ClassEntropy - EntropyAttr[g]

Node = max(Gain, key = Gain.get)

tree = insertNode(tree, addTo, Node)
for nD in Attr[Node]:
tree = insertNode(tree, Node, nD)
newData = data[data[Node] = = nD].drop(Node, axis = 1)
AttributeList=list(newData)[:-1] #New Attribute List
tree = getNextNode(newData, AttributeList, concept, conceptVals, tree, nD)
return tree

def main():
from pandas import DataFrame
data = DataFrame.from_csv('PlayTennis.csv')
print(data)
AttributeList = list(data)[:-1]
concept = str(list(data)[-1])
conceptVals = list(set(data[concept]))
tree = getNextNode(data, AttributeList, concept, conceptVals, {'root':'None'}, 'root')
print(tree)

main()

Output:
 10

Dataset:

Source: https://github.com/ggrao1/decision-tree
 11

Experiment 4: Artificial Neural Network using Backpropagation

Algorithm
Aim: Build an Artificial Neural Network by implementing the Backpropagation
algorithm and test the same using appropriate data sets.
import numpy as np
X = np.array(([2, 9], [1, 5], [3, 6]), dtype=float)
y = np.array(([92], [86], [89]), dtype=float)
X = X/np.amax(X,axis=0)
y = y/100

def sigmoid (x):

return 1/(1 + np.exp(-x))

def derivatives_sigmoid(x):
return x * (1 - x)

epoch=7000
learning_rate=0.1
inputlayer_neurons = 2
hiddenlayer_neurons = 3
output_neurons = 1

wh=np.random.uniform(size=(inputlayer_neurons,hiddenlayer_neurons))
bh=np.random.uniform(size=(1,hiddenlayer_neurons))
wo=np.random.uniform(size=(hiddenlayer_neurons,output_neurons))
bo=np.random.uniform(size=(1,output_neurons))

for i in range(epoch):
net_h=np.dot(X,wh) + bh
sigma_h= sigmoid(net_h)
net_o= np.dot(sigma_h,wo)+ bo
output = sigmoid(net_o)
deltaK = (y-output)* derivatives_sigmoid(output)
deltaH = deltaK.dot(wo.T) * derivatives_sigmoid(sigma_h)
wo = wo + sigma_h.T.dot(deltaK) *learning_rate
wh = wh + X.T.dot(deltaH) *learning_rate

print("Input: \n" + str(X))

print("Actual Output: \n" + str(y))
print("Predicted Output: \n" ,output)
 12

Output:

Source: https://github.com/praahas/machine-learning-vtu
 13

Experiment 5: Naïve Bayes Classifier

Aim: Write a program to implement the Naïve Bayes classifier for a sample
training data set stored as a .CSV file. Compute the accuracy of the classifier,
considering few test data sets.
def probAttr(data,attr,val):
Total=data.shape[0]
cnt = len(data[data[attr] == val])
return cnt,cnt/Total

def train(data,Attr,conceptVals,concept):
conceptProbs = {}
countConcept={}
for cVal in conceptVals:
countConcept[cVal],conceptProbs[cVal] = probAttr(data,concept,cVal)
AttrConcept = {}
probability_list = {}
for att in Attr: #Create a tree for attribute
AttrConcept[att] = {}
probability_list[att] = {}
for val in Attr[att]:
AttrConcept[att][val] = {}
a,probability_list[att][val] = probAttr(data,att,val)
for cVal in conceptVals:
dataTemp = data[data[att]==val]
AttrConcept[att][val][cVal] = len(dataTemp[dataTemp[concept] == cVal])/countConcept[cVal]

print("P(A) : ",conceptProbs,"\n")
print("P(X/A) : ",AttrConcept,"\n")
print("P(X) : ",probability_list,"\n")
return conceptProbs,AttrConcept,probability_list

def test(examples,Attr,concept_list,conceptProbs,AttrConcept,probability_list):
misclassification_count=0
Total = len(examples)
for ex in examples:
px={}
for a in Attr:
for x in ex:
for c in concept_list:
if x in AttrConcept[a]:
if c not in px:
px[c] = conceptProbs[c]*AttrConcept[a][x][c]/probability_list[a][x]
else:
px[c] = px[c]*AttrConcept[a][x][c]/probability_list[a][x]
print(px)
classification = max(px,key=px.get)
print("Classification :",classification,"Expected :",ex[-1])
if(classification!=ex[-1]):
misclassification_count+=1
misclassification_rate=misclassification_count*100/Total
accuracy=100-misclassification_rate
 14

print("Misclassification Count={}".format(misclassification_count))
print("Misclassification Rate={}%".format(misclassification_rate))
print("Accuracy={}%".format(accuracy))

def main():
import pandas as pd
from pandas import DataFrame
data = DataFrame.from_csv('PlayTennis_train1.csv')
concept=str(list(data)[-1])
concept_list = set(data[concept])
Attr={}
for a in list(data)[:-1]:
Attr[a] = set(data[a])
conceptProbs,AttrConcept,probability_list = train(data,Attr,concept_list,concept)

examples = DataFrame.from_csv(PlayTennis_test1.csv')
test(examples.values,Attr,concept_list,conceptProbs,AttrConcept,probability_list)

main()

Output:
 15

Dataset:
Training Set

Testing example

Source: https://github.com/ggrao1/NaiveBayes
 16

Experiment 6: Naïve Bayes Classifier using API

Aim: Assuming a set of documents that need to be classified, use the naïve
Bayesian Classifier model to perform this task. Built-in Java classes/API can be
used to write the program. Calculate the accuracy, precision, and recall for your
data set.
import pandas as pd
msg = pd.read_csv('document.csv', names=['message', 'label'])
print("Total Instances of Dataset: ", msg.shape[0])
msg['labelnum'] = msg.label.map({'pos': 1, 'neg': 0})
X = msg.message
y = msg.labelnum
from sklearn.model_selection import train_test_split
Xtrain, Xtest, ytrain, ytest = train_test_split(X, y)
from sklearn.feature_extraction.text import CountVectorizer
count_v = CountVectorizer()
Xtrain_dm = count_v.fit_transform(Xtrain)
Xtest_dm = count_v.transform(Xtest)
df = pd.DataFrame(Xtrain_dm.toarray(),columns=count_v.get_feature_names())
print(df[0:5])
from sklearn.naive_bayes import MultinomialNB
clf = MultinomialNB()
clf.fit(Xtrain_dm, ytrain)
pred = clf.predict(Xtest_dm)

for doc, p in zip(Xtrain, pred):

p = 'pos' if p == 1 else 'neg'
print("%s -> %s" % (doc, p))

from sklearn.metrics import accuracy_score, confusion_matrix, precision_score, recall_score

print('Accuracy Metrics: \n')

print('Accuracy: ', accuracy_score(ytest, pred))
print('Recall: ', recall_score(ytest, pred))
print('Precision: ', precision_score(ytest, pred))
print('Confusion Matrix: \n', confusion_matrix(ytest, pred))
 17

Output:

Source: https://github.com/rumaan/machine-learning-lab-vtu
 18

Dataset:
 19

Experiment 7: Bayesian Network

Aim: 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. You can use Java/Python ML library classes/API.
from pgmpy.models import BayesianModel
cancer_model = BayesianModel([('Pollution', 'Cancer'),('Smoker', 'Cancer'),('Cancer', 'Xray'),('Cancer',
'Dyspnoea')])
cancer_model.nodes()
cancer_model.edges()
cancer_model.get_cpds()
from pgmpy.factors.discrete import TabularCPD

cpd_poll = TabularCPD(variable='Pollution', variable_card=2, values=[[0.9], [0.1]])

cpd_smoke = TabularCPD(variable='Smoker', variable_card=2, values=[[0.3], [0.7]])
cpd_cancer = TabularCPD(variable='Cancer', variable_card=2, values=[[0.03, 0.05, 0.001, 0.02],
[0.97, 0.95, 0.999, 0.98]],evidence=['Smoker', 'Pollution'], evidence_card=[2, 2])
cpd_xray = TabularCPD(variable='Xray', variable_card=2, values=[[0.9, 0.2], [0.1, 0.8]],
evidence=['Cancer'], evidence_card=[2])
cpd_dysp = TabularCPD(variable='Dyspnoea', variable_card=2, values=[[0.65, 0.3], [0.35, 0.7]],
evidence=['Cancer'], evidence_card=[2])
cancer_model.add_cpds(cpd_poll, cpd_smoke, cpd_cancer, cpd_xray, cpd_dysp)
cancer_model.check_model()

cancer_model.get_cpds()
print(cancer_model.get_cpds('Pollution'))
print(cancer_model.get_cpds('Smoker'))
print(cancer_model.get_cpds('Xray'))
print(cancer_model.get_cpds('Dyspnoea'))
print(cancer_model.get_cpds('Cancer'))

cancer_model.local_independencies('Xray')
cancer_model.local_independencies('Pollution')
cancer_model.local_independencies('Smoker')
cancer_model.local_independencies('Dyspnoea')
cancer_model.local_independencies('Cancer')
cancer_model.get_independencies()

from pgmpy.inference import VariableElimination

cancer_infer = VariableElimination(cancer_model)

q = cancer_infer.query(variables=['Cancer'], evidence={'Smoker': 1})

print(q['Cancer'])

q = cancer_infer.query(variables=['Cancer'], evidence={'Smoker': 1,'Pollution': 1})

print(q['Cancer'])
 20

Output:

Inferencing:
 21

Diagnosis of heart patients using standard Heart Disease Data Set:

import numpy as np
from urllib.request import urlopen
import urllib
import matplotlib.pyplot as plt # Visuals
import seaborn as sns
import sklearn as skl
import pandas as pd

Cleveland_data_URL = 'http://archive.ics.uci.edu/ml/machine-learning-databases/heart-
disease/processed.hungarian.data'
np.set_printoptions(threshold=np.nan) #see a whole array when we output it

names = ['age', 'sex', 'cp', 'trestbps', 'chol', 'fbs', 'restecg', 'thalach', 'exang', 'oldpeak', 'slope', 'ca', 'thal',
'heartdisease']
heartDisease = pd.read_csv(urlopen(Cleveland_data_URL), names = names) #gets Cleveland data

del heartDisease['ca']
del heartDisease['slope']
del heartDisease['thal']
del heartDisease['oldpeak']

heartDisease = heartDisease.replace('?', np.nan)

from pgmpy.models import BayesianModel

from pgmpy.estimators import MaximumLikelihoodEstimator, BayesianEstimator

model = BayesianModel([('age', 'trestbps'), ('age', 'fbs'), ('sex', 'trestbps'), ('sex', 'trestbps'),

('exang', 'trestbps'),('trestbps','heartdisease'),('fbs','heartdisease'),
('heartdisease','restecg'),('heartdisease','thalach'),('heartdisease','chol')])

# Learing CPDs using Maximum Likelihood Estimators

model.fit(heartDisease, estimator=MaximumLikelihoodEstimator)

print(model.get_cpds('age'))
print(model.get_cpds('chol'))
print(model.get_cpds('sex'))
model.get_independencies()

from pgmpy.inference import VariableElimination

HeartDisease_infer = VariableElimination(model)

q = HeartDisease_infer.query(variables=['heartdisease'], evidence={'age': 28})

print(q['heartdisease'])

q = HeartDisease_infer.query(variables=['heartdisease'], evidence={'chol': 100})

print(q['heartdisease'])
 22

Output:
Diagnosis:
 23

Experiment 8: Clustering using EM Algorithm & k-Means Algorithm

Aim: 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. You can add
Java/Python ML library classes/API in the program.
from sklearn.cluster import KMeans
from sklearn import preprocessing
from sklearn.mixture import GaussianMixture
from sklearn.datasets import load_iris
import sklearn.metrics as sm
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

dataset=load_iris()
X=pd.DataFrame(dataset.data)
X.columns=['Sepal_Length','Sepal_Width','Petal_Length','Petal_Width']
y=pd.DataFrame(dataset.target)
y.columns=['Targets']
plt.figure(figsize=(14,7))
colormap=np.array(['red','lime','black'])

#REAL PLOT
plt.subplot(1,3,1)
plt.scatter(X.Petal_Length,X.Petal_Width,c=colormap[y.Targets],s=40)
plt.title('Real')

#KMeans -PLOT
plt.subplot(1,3,2)
model=KMeans(n_clusters=3)
model.fit(X)
predY=np.choose(model.labels_,[0,1,2]).astype(np.int64)
plt.scatter(X.Petal_Length,X.Petal_Width,c=colormap[predY],s=40)
plt.title('KMeans')

#GMM PLOT
scaler=preprocessing.StandardScaler()
scaler.fit(X)
xsa=scaler.transform(X)
xs=pd.DataFrame(xsa,columns=X.columns)
gmm=GaussianMixture(n_components=3)
gmm.fit(xs)
y_cluster_gmm=gmm.predict(xs)
plt.subplot(1,3,3)
plt.scatter(X.Petal_Length,X.Petal_Width,c=colormap[y_cluster_gmm],s=40)
plt.title('GMM Classification')
 24

Output
 25

Dataset:
 26

Experiment 9: k-Nearest Neighbour Algorithm

Aim: Write a program to implement k-Nearest Neighbour algorithm to classify
the iris data set. Print both correct and wrong predictions. Java/Python ML library
classes can be used for this problem.
from sklearn.datasets import load_iris
from sklearn.neighbors import KNeighborsClassifier
from sklearn.model_selection import train_test_split
import numpy as np
dataset=load_iris()
X_train,X_test,y_train,y_test=train_test_split(dataset["data"],dataset["target"],random_state=0)
clf=KNeighborsClassifier(n_neighbors=1)
clf.fit(X_train,y_train)
for i in range(len(X_test)):
x=X_test[i]
x_new=np.array([x])
prediction=clf.predict(x_new)

print("TARGET=",y_test[i],dataset["target_names"][y_test[i]],"PREDICTED=",prediction,dataset["target_
names"][prediction])
print(clf.score(X_test,y_test))

Output:
TARGET= 2 virginica PREDICTED= [2] ['virginica'] TARGET= 0 setosa PREDICTED= [0] ['setosa']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 2 virginica PREDICTED= [2] ['virginica']
TARGET= 0 setosa PREDICTED= [0] ['setosa'] TARGET= 1 versicolor PREDICTED= [1] ['versicolor']
TARGET= 2 virginica PREDICTED= [2] ['virginica'] TARGET= 0 setosa PREDICTED= [0] ['setosa']
TARGET= 0 setosa PREDICTED= [0] ['setosa'] TARGET= 0 setosa PREDICTED= [0] ['setosa']
TARGET= 2 virginica PREDICTED= [2] ['virginica'] TARGET= 2 virginica PREDICTED= [2] ['virginica']
TARGET= 0 setosa PREDICTED= [0] ['setosa'] TARGET= 0 setosa PREDICTED= [0] ['setosa']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 0 setosa PREDICTED= [0] ['setosa']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 1 versicolor PREDICTED= [1] ['versicolor']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 1 versicolor PREDICTED= [1] ['versicolor']
TARGET= 2 virginica PREDICTED= [2] ['virginica'] TARGET= 0 setosa PREDICTED= [0] ['setosa']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 2 virginica PREDICTED= [2] ['virginica']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 1 versicolor PREDICTED= [1] ['versicolor']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 0 setosa PREDICTED= [0] ['setosa']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 2 virginica PREDICTED= [2] ['virginica']
TARGET= 0 setosa PREDICTED= [0] ['setosa'] TARGET= 2 virginica PREDICTED= [2] ['virginica']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 1 versicolor PREDICTED= [1] ['versicolor']
TARGET= 1 versicolor PREDICTED= [1] ['versicolor'] TARGET= 0 setosa PREDICTED= [0] ['setosa']
TARGET= 0 setosa PREDICTED= [0] ['setosa'] TARGET= 1 versicolor PREDICTED= [2] ['virginica']

0.973684210526
 27

Dataset:

Source: https://github.com/praahas/machine-learning-vtu
 28

Experiment 10: Locally Weighted Regression Algorithm

Aim: Implement the non-parametric Locally Weighted Regression algorithm in
order to fit data points. Select appropriate data set for your experiment and draw
graphs.
from math import ceil
import numpy as np
from scipy import linalg

def lowess(x, y, f, iterations):

n = len(x)
r = int(ceil(f * n))
h = [np.sort(np.abs(x - x[i]))[r] for i in range(n)]
w = np.clip(np.abs((x[:, None] - x[None, :]) / h), 0.0, 1.0)
w = (1 - w ** 3) ** 3
yest = np.zeros(n)
delta = np.ones(n)
for iteration in range(iterations):
for i in range(n):
weights = delta * w[:, i]
b = np.array([np.sum(weights * y), np.sum(weights * y * x)])
A = np.array([[np.sum(weights), np.sum(weights * x)],[np.sum(weights * x), np.sum(weights * x *
x)]])
beta = linalg.solve(A, b)
yest[i] = beta[0] + beta[1] * x[i]

residuals = y - yest
s = np.median(np.abs(residuals))
delta = np.clip(residuals / (6.0 * s), -1, 1)
delta = (1 - delta ** 2) ** 2

return yest

def main():
import math
n = 100
x = np.linspace(0, 2 * math.pi, n)
y = np.sin(x) + 0.3 * np.random.randn(n)
f =0.25
iterations=3
yest = lowess(x, y, f, iterations)

import matplotlib.pyplot as plt

plt.plot(x,y,"r.")
plt.plot(x,yest,"b-")

main()
 29

Output: