HEART DISEASE PROGNOSIS WITH ARTIFICIAL

INTELLIGENCE AND NEURAL NETWORKS

A PROJECT REPORT - II

Submitted by

J. JENI NENATHA (212922405001)

In partial fulfilment for the award of the degree

of

MASTER OF ENGINEERING

in

COMPUTER SCIENCE AND ENGINEERING

ST. JOSEPH COLLEGE OF ENGINEERING, SRIPERUMBUDUR

ANNA UNIVERSITY: CHENNAI 600025

JULY 2024
 ANNA UNIVERSITY: CHENNAI 600 025

BONAFIDE CERTIFICATE
Certified that this project report “HEART DISEASE PROGNOSIS WITH
ARTIFICIAL INTELLIGENCE AND NEURAL NETWORKS”is the bonafide
work of “J. JENI NENATHA (212922405001)” who carried out theproject work
under my supervision.

SIGNATURE SIGNATURE

Dr. M. NAVANEETHA KRISHNAN Dr. M. NAVANEETHA KRISHNAN

SUPERVISOR HEAD OF THE DEPARTMENT

Department of Computer Science Department of Computer Science
and Engineering. and Engineering.
St. Joseph College of Engineering, St. Joseph College of Engineering,
Sriperumbudur, Chennai- 602117. Sriperumbudur, Chennai- 602117.

Submitted for the Master of Engineering Degree Viva- Voce held at

St. Joseph College of Engineering on ………………

INTERNAL EXAMINER EXTERNAL EXAMINER

 ACKNOWLEDGEMENT

We are very grateful and gifted in taking up this opportunity to thank the
Lord Almighty for showering his unlimited blessings upon us.

We take this occasion to express our respect and thanks to Rev. Fr. Dr. J. E.
Arul Raj, Founder and Chairman and Rev. Sr. S. Gnanaselvam DMI, Managing
Trustee, for facilitating us to do our project successfully.

We take this opportunity to thank our administrator Rev. Fr. L. Savariappan

MMI, for his kind and encouraging guidance.

We wish to eloquent our genuine thanks to Dr. T. Ahilan M.E., Ph.D.,Principal

for his support and guidance.
We express our deep sense of gratitude to HOD & Project Guide Dr. M.
Navaneetha Krishnan, M.E., Ph.D., for his scintillating discussions and
encouragement towards our project work.

It is pleasure to acknowledge our indebtedness to all the Staff Members of our

department who aided us ingeniously to bring our project as effective one. Our special
thanks to Non-Teaching Staffs for extending the Lab facilities.

We thank our family members and friends for their honorable support.
 ABSTRACT

Mortality rate increases all over the world on daily basis. The reasons for

this could be increase in the numbers of patient with cardiovascular disease. When

considering death rates and large number of people who suffers from heartdisease,

it is revealed how important early diagnosis of heart disease. Traditional way of

predicting heart disease is doctor’s examination or number of medical tests such

as ECG, Stress Test, and Heart MRI etc. Nowadays, Health care industry contains

huge amount of heath care data, which contains hidden information. This hidden

information is useful for making effective decisions. Computer based information

along with advanced Data mining techniques are used for appropriate results.

Neural network is widely used tool for predicting heart disease diagnosis. In our

project, a heart disease prediction system which uses artificial neural network

backpropagation algorithm is proposed. 14 clinical features were used as input for

the neural network and then the neural network was trained with back

propagationalgorithm to predict absence or presence of heart disease.

i
 TABLE OF CONTENT

CHAPTER NO. TITLE PAGE NO.

ABSTRACT iv
LIST OF ABBREVIATIONS vii
LIST OF FIGURES viii
1 INTRODUCTION 01
1.1 OVERVIEW OF DOMAIN 01
1.2 CARDIOVASCULAR RISK FACTOR 03
1.3 BACKGROUND 04
1.4 ARTIFICIAL NEURAL NETWORKS 05
1.5 HEART DISEASE 07
1.6 PROBLEM STATEMENT 08
1.7 OBJECTIVE 09
2 LITERATURE SURVEY 10
3 SYSTEN ANALYSIS 18
3.1 EXISTING SYSTEM 18
3.1.1 Disadvantages 20
3.2 PROPOSED SYSTEM 20
3.3 ARTIFICIAL NEURAL NETWORK 21
3.4 ADVANTAGES 23
4 SYSTEM DESIGN 24
4.1 SYSTEM ARCHITECTURE 24
4.2 HEART DISEASE PREDICTION 24
4.3 MODULES 25
4.3.1 Data Collection 25
4.3.2 Training and Testing Dataset 27
4.3.3 Building Neural Network 28
4.3.4 Performance Measures 29
5 SYSTEM DESCRIPTION 31

ii
 5.1 HARDWARE DESCRIPTION 31
5.2 SOFTWARE DESCRIPTION 31
5.2.1 Python Programming 32
5.2.2 Jupyter Notebook 34
6 SYSTEM IMPLEMENTATION 36
6.1 SOFTWARE IMPLEMENTATION 37
6.1.1 Implementation Methods 37
6.1.2 Import Libraries 38
6.1.3 Analyze and Visualize Pandas 39
Data Structure
6.1.4 Exploratory Data Analysis 40
6.1.5 Checking Relationship Among 40
Variables
6.1.6 Checking and Dropping 40
Duplicates
6.1.7 Outlier Detection and Removal 40
6.1.8 Pre-processing 41
6.1.9 Fitting an ANN Model 42
6.1.10 Confusion Matrix 42
6.1.11 ANN Model Classification and 42
Saving
7 SYSTEM TESTING 43
7.1 UNIT TESTING 43
7.2 VALIDATION TESTING 45
7.3 MODULE LEVEL TESTING 45
7.4 INTEGRATION & SYSTEM TESTING 46
7.5 REGRESSION TESTING 46
8 RESULTS AND DISCUSSION 47
8.1 MODULES IMPORT 47
8.2 DATASET VISUALIZATION 49
iii
 8.3 CORRELATION MATRIX 50
8.4 FEATURE SELECTION 50
8.5 ANN MODEL TRAINING 51
8.6 PREDICTION 52
8.7 CLASSIFICATION 52
8.8 PERFORMANCE EVALUATION 54
9 CONCLUSION 55
REFERNCES 56

iv
 LIST OF ABBREVIATIONS

ACO - Ant Colony Optimization

ANN - Artificial Neural Networks
CRT - Classification and Regression Tree
CVD - Cardio Vascular Disease
FSBFS - Fast Correlation-Based Feature Selection
IHDPS - Intelligent Heart Disease Prediction System
KNN - K- Nearest Neighbor
PSO - Particle Swarm Optimization
SVM - Support Vector Machine

v
 LIST OF FIGURES

FIGURE NO. TITLE PAGE NO.

1.1 Example of Artificial Neural Network 07

3.1 Example ANN 22

4.1 System Architecture 24

vi
 CHAPTER 1

INTRODUCTION

1.1 OVERVIEW OF DOMAIN

Heart is made up of various Nerves and muscles. Any failure or defect

of heart may lead to sudden death. Nowadays, in the world, heart disease is the
major cause of deaths. The World Health Organization (WHO) has estimated that 12
million deaths occur worldwide, every year due to the heart diseases. In 2008,
17.3 million people died due to heart disease. Over 80% of deaths in world are
because of heart disease. WHO estimated by 2030, almost 23.6 million people
will die due to heart disease. This is one of the reasons why researcher has focus
more in designing intelligent system that can be used to diagnose heart diseases
with high accuracy, to avoid misdiagnosis. Besides, many people are living with
heart disease without awareness. If heart disease could be predicted before, lots
of patient deaths would be prevented and also a more accurate and efficient
treatment way could be provided. Predication should be done to reduce risk of
heart disease. Diagnosis is usually based on signs, symptoms and physical
examination of a patient. Almost all the doctors are predicting heart disease by
learning and experience. Developing a medical diagnosis system based on
machine learning for prediction of heart disease provides more accurate diagnosis
than traditional way and reduces cost of treatment. In our project, prediction of
heart disease by an automated medical diagnosis system based on machine
learning is proposed to satisfy this need. Backpropagation Algorithm which is
commonly used Artificial Neural Network learning methodology was used for the
prediction system. This system can help in diagnosing disease with less medical
tests & effective treatments.

Heart Disease is one of the leading diseases around the world. The effective
functioning of the heart plays a vital role in the body. There are many types of

1
heart diseases such as Myocardial infarction, Myocardial ischemia, Congenital
heart disease, Coronary heart disease, Cardiac arrest, Peripheral heart disease etc.,
There are various methodologies available in predicting the heart disease.

Machine learning is a type of artificial intelligence that makes the machines

to learn from training data and makes predictions on the test data based on the
learned data. The basic idea behind the machine learning is to find the patterns
among the data and make the predictions. There are numerous applications in
machine learning such as in recommender systems, medical diagnosis,
bioinformatics etc., Basically, there are three types of learning in machine
learning such as supervised learning, unsupervised learning and reinforcement
learning.

Heart disease describes a range of condition that affects the heart. The term
cardiac disease is regularly utilized with cardio vascular disease (CVD). The
blood to the heart is supplied by coronary supply routes and narrowing of
coronary arteries is the major cause for heart failure. Prediction of cardiovascular
disease is considered as one of the most important subjects in the section of data
analytics. The major cause of heart attack in United States is coronary artery
disease. Cardiac disorder is widespread in male than that of female. The survey
analyzed by World Health Organization (WHO) estimates that 24% of people died
in India due to cardiac disorder. Analysts have recorded the differentcomponents
that increment the possibility of cardiac disorder and coronary arterydisease.

The elements are put down into two types that are the elements that could
be modified and elements that could not be modified. Age, smoking, heredity,
sex, high blood pressure, poor diet, consumption of alcohol, physical inactivity is
considered to be the risk factor of heart diseases. Age, sex, heredity are the
elements that could not be modified while the elements like smoking, intake of

2
alcohol can be modified. The determinant condition related to lifestyle could be
vicissitude with meditation and lifestyle changes. The most common method used
for treating coronary artery disease was angiography. Angiography is the costliest
method and have more reactions. The risk factors make the detection of heart
diseases very difficult; these constrains leads to the evolution of modern methods
which detects the presence of cardiac diseases. It is hard to identify cardiac
disorders manually hence conventional methods are used which is mostly based
on examination of patient data. Machine learning techniques are useful for
predicting the diseases. We develop a diagnosis system which is based on  2 -
DNN. The most objective of the determination framework is to dispense with the
issue of over fitting and under fitting. The training data is feed in to the neural
organizes. The testing dataset is utilized to access the execution of the neural
organize. The DNN with multiple hidden layers is used hence the proposed model
has high performance than ANN.  2 -DNN is developed for improving the
classification accuracy of cardiac disorder prognosis.

1.2 CARDIOVASCULAR RISK FACTOR

A cardiovascular risk factor is defined as a factor in which the patient's

exposure to this factor increases the risk of developing a CVD, while the removal
or improvement of this factor decreases this risk. The importance of the risk factor
considered is defined by the strength of the association with the disease
(expressed by the relative risk observed in the exposed subjects compared to the
unexposed), and or the gradual association (parallel to the level of the risk factor).
This definition implies a notion of causality between the factor and the disease.
This means that the risk factor precedes the disease (notion of anteriority) and that the
correction of the factor must make it possible to reduce the incidence of the
disease (concept of reversibility). It must of course be recognized in several
different populations and offer a plausible physio pathological explanation of the
disease. Strictly speaking, when there is no direct causal relationship, it is actually

3
a "risk marker", a witness to a process (e.g., elevation of microalbuminuria,
elevation of C-reactive protein "CRP"). While there is a direct causal link between
the agent and the disease, it is a genuine risk factor. Overall cardiovascular risk
is defined as the probability of developing a cardiovascular event in a given time
depending on several risk factors considered simultaneously. Born more than 50
years ago with the advent of the analytical epidemiology of cardiovascular
diseases, this concept now appears as a potential prevention tool integrating the
multifactorial nature of these diseases.

A. Physiological Factors
1) Age
2) Sex
3) Menopausal Status
B. Lifestyle Factors

1) Smoking
2) Physical Activity
3) Alcohol

1.3 BACKGROUND

Numerous frameworks for coronary disorder forecast utilizing AI

calculation were developed early. Algorithm like back vector machine, K-nearest
neighbour and Artificial Neural Network was developed early to anticipate the
existence or nonappearance of cardiac disorder. Survey indicates that ANN based
models are mostly used in heart disease prediction and the accuracy of previous
works was less in contrast with the developed model. S. Radhimeenakshi
proposed a classification system for the prediction of cardiac disorder. The work
which he carried out focuses on two algorithms namely Support Vector Machine
(SVM) and Decision Tree. He carried out his work in weak and python the same
predictive model is used for both the training dataset and the testing dataset. The

4
model achieved more accuracy when decision tree classification algorithm is
used. The accuracy of the model is evaluated using confusion matrix and the
model achieved accuracy of 55% for SVM.

1.4 ARTIFICIAL NEURAL NETWORKS

The Artificial Neural Networks (ANN) additionally named as a neural

organization is a numerical model dependent on natural neural organizations. An
artificial neural organization depends upon the perception for cerebrum of a
human. The mind of human is a convoluted trap of neurons. The similar fake
neural organization is a connected arrangement of straightforward components to
be specific info, covered up, and yield component. The credits which are accepted
as contributions to the following structure. In clinical assessment of patients,
hazard factors are treated as contributions to counterfeit neural organization.
The natural occurrences are generally utilized in the research field by the
researchers due to the versatility and stimulated for developing the intelligent
optimization algorithm. Due to the easiness and simplification, Particle Swarm
Optimization (PSO) algorithm is effectively used to heart disease prediction. PSO
is simply used in the optimal solution for the difficulties in the prediction of heart
disease. The ACO algorithm is introduced with the combination of the
optimization to enhance the performance of the model. It is also established to
provide the solution for the difficulties in optimization. The decision variable
provides continuous domain which is characteristic by the problem defined in the
heart disease prediction. The single optimization algorithm has some drawbacks
such as low efficiency and generalization in providing solution for complex
issues. The application of intelligent optimization is explored in biotechnology
for determining the illness of the patient. The exploitation and exploration of the
binary and multi-class heart disease prediction is obtained by combining PSO and
ACO technique. The redundant and irrelevant features are obtained by utilizing
the Fast Correlation-Based Feature selection (FCBF) method and the outcome of

5
PSO is obtained by initial rate of ACO. After the parameter changes, the
classification of heart disease model is created. The best algorithm for the
classification and heart disease prediction is K-Nearest Neighbour and the
Random Forest algorithm.

Famous neural network algorithm is Hopfield, Multilayer insight, counter

engendering networks, outspread premise capacity, and self-putting together
guides, and so on. The forward neural system organization is the major and least
complex sort of fake neural organization comprises of input layer, hidden layer,
and output layer. The neural organization should be an arranged to create the ideal
arrangement of the output.

There are three learning method in neural network,

1) Supervised Learning,
2) Unsupervised Learning,
3) Reinforcement learning

The perception is an essential tool of an Artificial Neural Network which

is utilized for order in which the examples are directly distinct. The fundamental
technique of neuron utilized in insight is the McCulloch-Pitts technique.

The perception receipts an info esteem vector and yields 1 for the outcome
which is more noteworthy than predefined edges and -1 in any case. The
verification assembly of calculation is called as the apperception combination
hypothesis. Figure 1.1 shows the example of ANN. The yield hub isutilized to
speak to the system which provide the nodes in neural organization engineering
is ordinarily called neurons. Every info node is associated with the yield hub by
means of connection. This is utilized for copying the strong point ofthe synaptic
association among neurons.

6
 Fig 1.1: Example of Artificial Neural Network

1.5 HEART DISEASE

Coronary heart disease happens occur when the supply routes of the heart
typically give oxygen and blood to the heart is limited and totally hindered.
Cardiovascular sicknesses represent huge death and grimness in world. In India,
death because of Chewier 1.6 million in the year 2000, 201561 million cases will
be because of coronary heart disease.

The hazard features that cause heart disease are

1) Blood pressure
2) Cholesterol
3) Diabetes
4) Smoking
5) Eating Habits
6) Sedentary life style
7) Stress

7
 This pandemic might be stopped through the advancement of better ways
of life, active work, conventional food utilization would assist with alleviating
this weight.

1.6 PROBLEM STATEMENT

The use of machine learning methods for classification and prediction of

heart disease have examined in previous research studies. The studies are
emphasis on the specific effects of the machine learning methods and not on the
optimization technique used for optimization. The hybrid optimization methods
are utilized in few researchers’ challenges for the optimized organization of
machine learning. The optimization technique namely Ant Colony Optimization
and Particle Swarm Optimization are the proposed studies of the literature which
is specified with ML technique namely Support Vector Machine (SVM), K-
Nearest Neighbour (KNN) and Random Forest.

The Fast Correlation-Based Feature Selection (FCBF) technique is utilized

in the proposed work. The attributes are selected relevant to mining between the
original attributes when all the continuous attributes are discretized. The feature
selection is used in the pre-processing step in machine learning which provide
efficient in decreasing dimensionality, removal of unwanted data, improving the
learning efficiency and increasing result of proposed technique. Artificial Neural
Network is applied in the second step which is utilized to select the relevant
features of dataset. The characteristic selection techniques select the best subset
characteristics which improves the efficiency of the classification. Hence,
classification method is applying in third step to predict heart disease and
classification efficiency is measured to estimate the performance characteristic of
selection methods.

8
1.7 OBJECTIVE

The prediction of heart disease using various classification algorithm like

Support Vector Machine, K-Nearest Neighbour, Naïve Bayes, Multilayer
Perception and Random Forest along with Artificial Neural Network is the main
objective of the proposed project. Ant Colony Optimization (ACO) combined
with Particle Swarm Optimization (PSO) method optimize Artificial Neural
Network. The data in the heart disease is analysed using weak data-mining tool.

The main assistances of this project are:

1. Extracting of categorized efficiency is beneficial for heart disease

diagnosing.
2. Fast Correlation-Based Feature selection (FCBF) approach for elimination
of unnecessary and unrelated features.
3. Features is optimized with Particle Swarm Optimization (PSO) then
outcome of PSO is assigned as the original standards of Ant Colony
Optimization (ACO) method.
4. The different data mining methods is comparison on the heart disease
dataset.
5. Identifying the best performance-based technique for heart disease
estimation.

9
 CHAPTER 2
LITERATURE SURVEY

B. Narasimhan and Dr. A. Malathi (2019) Altered particle swarm

optimization-based attribute selection strategy with improved fuzzy
Artificial Neural Network classifier for coronary artery heart disease risk
prediction [1]

Their project explains the concept of the use of delicate figuring in choice
emotionally supportive network in sickness analysis is one of the creating
interdisciplinary exploration territories in the field of software engineering. AI
calculations assumes a significant part in danger location of sicknesses. Highlight
determination between the dataset is the fundamental factor that impacts coronary
illness forecast exactness. Mathew's connection coefficient execution metric is
additionally considered. Molecule swarm streamlining calculation is changed and
pragmatic for execution quality determination. Upgraded fluffy counterfeit neural
organization classifier plays out the forecast task.

Manikandan. T, Kumar and Vinoth (2019) discovered Heart Disease

Prediction System Using Artificial Neural Network, Radial Basis Function
and Case Based Reasoning [2]
Their project explains the concept of Heart disease which is perhaps the most
hazardous sicknesses to human which shows the best approach to death
everywhere on the world for a very long time. Numerous examinations have been
finished with the techniques for information recognition in a few fields for heart
disease expectation and have demonstrated the standard degrees of precision. By
looking at the audit of those effectiveness levels, their exploration project is
recommended to help specialists not exclusively to recognize and anticipate the
coronary illness by accomplishing accuracy levels yet additionally assists with
proposing the medication effectively as indicated by the anticipated infection. In
10
the project valuation is finished by two strategies, for example, ANN (Artificial
neural organization) by testing the datasets and CBR (Case Based Reasoning)
picture likeness search by planning the similarities of pictures of old patients kept in
information base for the assessment of coronary illness. The outcome in valuation
of CBR is likewise executed for suggesting medication from the past ofold patients
with Generalized Regression Neural Network and Radial premise work viably.

Annisa Darmawahyuni, Siti Nurmaini (2019) Coronary Heart Disease

Interpretation Based on Deep Neural Network [3]

Their project explains the idea of coronary heart disease (CHD) builds each
year by a critical amount of morality. In addition, death from coronary heart
disease acquires the most elevated predominance in Indonesia at 1.5 percent. The
misdiagnosis of coronary heart disease is an urgent central which is the main
consideration which create demise. To avoid wrong diagnosis of CHD the
insightful framework has planned. Their method is suggested a simulation that
can utilized for the analyze of coronary heart disease in preferable execution over
conventional indicative strategies. A few analysts have built up a framework
utilizing ordinary neural network and machine learning calculation. Yet,
outcomes do not provide a decent execution. In light of a regular neural
organization, deep neural organization (DNN) is suggested in work. The
supervised learning neural network provides the calculation that great in the
classification. In DNN model, the usage of double grouping was executed to
analyze CHD present or CHD missing. To help execution examination utilizing
the UCI machine archive coronary heart disease dataset, ROC Curve and its
disarray grid was actualized in the proposed method. The overall accuracy,
sensitivity, and specificity acquired for the proposed method is 96%, 99%, 92%.

11
Atiqur Rahman and Aurangzeb Khan (2019) An Automated Diagnostic
System for Heart Disease Prediction Based on _2 Statistical Model and
Optimally Configured Deep Neural Network [4]

Their project explains the concept of Different robotized choice emotionally

supportive networks dependent on Artificial Neural Network (ANN) have been
generally proposed for the recognition of coronary illness in past investigations.
In any case, the greater part of these methods centre around the pre-preparing of
highlights as it were. In their project, refinement of highlights and disposal of the
issues presented by the prescient model, i.e., the issues of underfitting and
overfitting. By evading the model from overfitting and underfitting, it can show
great execute organization arrangement and superfluous highlights regularly bring
about overfitting the preparation information. To dispose of unimportant
highlights, they propose to utilize _2 statistical model while the ideally designed
deep neural network (DNN) is looked by utilizing thorough exhausted stratergy.
The strength of the proposed half and half model named 2-DNN is assessed by
contrasting its exhibition and ordinary ANN and DNN models, another best-in
class AI models and recently announced techniquesfor heart disease expectation.
The proposed model accomplishes the expectation precision of 93.33%. The
acquired outcomes are promising contrasted with the recently revealed strategies.
The discoveries of the investigation recommend that the proposed symptomatic
framework can be utilized by doctors to precisely anticipate heart disease. The
proposed 2-DNN model demonstrates enhanced robustness against both
overfitting and underfitting by effectively filtering out irrelevant features. This
approach not only improves model reliability but also ensures that the system
remains adaptable to diverse datasets and clinical scenarios. Consequently, the 2-
DNN model represents a significant advancement in predictive analytics for heart
disease, offering valuable support for medical decision-making.

12
Bin Xiao, Yunqiu Xu, Xiuli Bi and Junhui Zhang (2019) discovered Heart
sounds classification using a novel 1-D convolutional neural network with
extremely low parameter consumption [5]
Their project explains the concept of Automatic heart sound is one of the normally
utilized procedures for cardiovascular infections discovery. A novel heart sound
arrangement technique dependent on profound learning advances for
cardiovascular infection expectation is presented, which is fundamentally
included three sections: pre-preparing, 1-D waveform heart sound patches
characterization utilizing a profound convolutional neural organization (CNN)
with consideration component, and larger part deciding in favor of conclusive
forecast of heart sound accounts. To upgrade the data stream of the CNNs, a
square stacked style design with coterie blocks is utilized, and in every faction
block a bidirectional association structure is presented in the proposed CNN. By
utilizing the stacked faction and change obstructs, the proposed CNN
accomplishes both spatial and channel consideration driving a promising
characterization execution. Also, a novel detachable convolution with upset
bottleneck is used to decouple the spatial and channel-wise significance of
highlights proficiently. Analyses on Physio Net/CinC 2016 show that the
proposed strategy acquires a predominant grouping result and dominates in
utilization of boundary contrasting with best-in-class strategies.

C. Beulah Christalin Latha and S. Carolin Jeeva (2019) discovered

Improving the accuracy of prediction of heart disease risk based on ensemble
classification techniques [6]

Their project explains the concept of Machine learning comprises of man-

made consciousness which is utilized in tackling several problems in information
science. The regular use of AI is the forecast of outcome dependent on previous
information. The machine takes in examples from the current dataset, and
afterward applies them to an obscure dataset to foresee the result. Order is an
13
incredible AI method that is ordinarily utilized for expectation. Some
characterization calculations foresee with acceptable exactness, while others
show a restricted precision. Their project explores a strategy named outfit
grouping, which is utilized for improving the exactness of powerless calculations
by joining numerous classifiers. Examinations with their device were performed
utilizing a coronary illness dataset. A similar logical methodology was done to
decide how the outfit procedure can be applied for improving forecast precision
in coronary illness. The focal point of their project isn't just on expanding the
exactness of powerless order calculations, yet additionally on the execution of the
calculation with a clinical dataset, to show its utility to foresee illness at a
beginning phase. The consequences of the investigation show that troupe
methods, for example, sacking and boosting are powerful in improving the
forecast precision of feeble classifiers, and display acceptable execution in
recognizing danger of coronary illness. A most extreme increment of 7%
precision for powerless classifiers were accomplished with the assistance of outfit
characterization. The presentation of the cycle was additionally upgraded with an
element choice usage, and the outcomes indicated huge improvement in forecast
precision.

Mohamed Elnoamany, Ashraf Dawood, Nahed Mohamed Momtaz, Waleed

Abdou, “Thrombospondin-1 Levels in Patients with Coronary HeartDisease”
WJCD (2021) [7]

Thrombospondin-1 (TSP1) is associated with atherosclerosis in animals with

diabetes mellitus (DM), but the precise role of TSP-1 in human atherosclerosis
remains unknown. Objectives: To investigate serum throm- bospondin1 level in
patients with coronary artery disease with and without type 2 DM and its
relationship to coronary artery scoring systems. Methods: The study of 180
patients recruited from those underwent coronary angiography for suspected
coronary artery disease (CAD) was approved by Institutional Review Board and
Institutional Ethical Committee for Human Research of Menoufia university
14
hospital. They were divided according to presence of CAD and type 2 DM into 4
groups: Group I (n = 44 patients): Non diabetic subjects without CAD, Group II
(n = 40 patients): Diabetic patients without CAD, Group III (n = 49 patients): Non
diabetic patients with CAD and Group IV (n = 47 patients): Diabetic patientswith
CAD. Serum level of TSP-1 was measured in all groups and coronary artery
scoring analysis was done. Results: Serum TSP-1 levels were higher in patients
with CADand DM than in other groups (P < 0.01) .

Pahulpreet Singh Kohli, Shriya Arora, “Application of Machine Learning in

Disease Prediction”, IEEE (2019) [8]

The application of machine learning in the field of medical diagnosis is

increasing gradually. This can be contributed primarily to the improvement in the
classification and recognition systems used in disease diagnosis which is able to
provide data that aids medical experts in early detection of fatal diseases and
therefore, increase the survival rate of patients significantly. In their work, they
apply different classification algorithms, each with its own advantage on three
separate databases of disease (Heart, Breast cancer, Diabetes) available in UCI
repository for disease prediction. The feature selection for each dataset was
accomplished by backward modeling using the p-value test. The results of the
study strengthen the idea of the application of machine learning in early detection
of diseases.

Bin Deng, Min Guo "Risk Factors and Intervention Status of Cardiovascular
Disease in Elderly Patients with Coronary Heart Disease", Health( 2020) [9]

Recent days, heart ailments assume a fundamental role in the world. The
physician gives different names for heart disease, for example, cardiovascular
failure, heart failure and so on. Among the automated techniques to discover the
coronary illness, this research work uses Named Entity Recognition (NER)
algorithm to discover the equivalent words for the coronary illness content to

15
mine the significance in clinical reports and different applications. The Heart
sickness text information given by the physician is taken for the preprocessing
and changes the text information to the ideal meaning, at that point the resultant
text data taken as input for the prediction of heart disease. This experimental work
utilizes the NER to discover the equivalent words of the coronary illness text data and
currently uses the two strategies namely Optimal Deep Learning and Whale
Optimization which are consolidated and proposed another strategy Optimal Deep
Neural Network (ODNN) for predicting the illness. For the prediction, weights
and ranges of the patient affected information by means of chosen attributes are
picked for the experiment. The outcome is then characterized with the Deep
Neural Network and Artificial Neural Network to discover the accuracy of the
algorithms. The performance of the ODNN is assessed by means for classification
methods, for example, precision, recall and f-measure values.

T. Velmurugan, U. Latha "Classifying Heart Disease in Medical Data Using

Deep Learning Methods", Journal of Computer and Communications(2021)
[10]

Data mining is a subfield in the subject of software engineering. It is the

methodical procedure of finding examples in huge data sets including techniques
at the crossing point of manufactured intelligence, machine learning, insights, and
database systems. The goal of the data mining methodology is to take data from
a data set and change it into a reasonable structure for further use. Our
examination concentrates on this part of Medical conclusion learning design
through the gathered data of diabetes and to create smart therapeutic choice
emotionally supportive networks to help the physicians. The primary target of this
examination is to assemble an Intelligent Diabetes Disease Prediction Systemthat
gives analysis of diabetes malady utilizing diabetes patient's database. In this
system, we propose the use of algorithms like Bayesian and KNN (K-Nearest

16
Neighbor) to apply on a diabetes patient’s database and analyze them by taking
various attributes of diabetes for prediction of diabetes disease.

P.Santhi a, R.Ajayb"A Survey on Heart Attack Prediction UsingMachine

Learning" Turkish Journal of Computer and Mathematics Education (2021)
[11]
Machine Learning Algorithm is used for many different diseases. Machine
Learning is a learning of machine by own itself. And it is a part of AI that deals
with to learn a machine according to their own. Now-a-days most are affected due
to Heart attack it becomes headache for doctors. In order to reduce the count of
death we need to predict the Heart attack. For this problem Machine Learning play
a major role in their project. This prediction takes a people from the danger zone
of their life. In their project we use KNN algorithm and Random forest algorithm.

17
 CHAPTER 3

SYSTEM ANALYSIS

3.1 EXISTING SYSTEM

Tremendous works in literature related with heart disease diagnosis using

data mining techniques have motivated their work. The researchers in the medical
field diagnose and predict the diseases in addition to providing effective care for
patients by employing the data mining techniques. The data mining techniques
have been employed by numerous works in the literature to diagnose diverse
diseases, for instance: Diabetes, Hepatitis, Cancer, Heart diseases and more A
model Intelligent Heart Disease Prediction System (IHDPS) built with the aid of
data mining techniques like Decision Trees, Naïve Bayes and Neural Network.
The problem of identifying constrained association rules for heart disease
prediction was studied by Carlos Ordonez. The assessed data set encompassed
medical records of people having heart disease with attributes for risk factors,
heart perfusion measurements and artery narrowing be removed before mining
process occurs. Association rule mining is a major data mining technique, and is
a most commonly used pattern discovery method. It retrieves all frequent patterns
in a data set and forms interesting rules among frequent patterns. Most frequently
used association rule mining methods are Apriority and Growth. Frequent Item
set Mining (FIM) is considered to be one of the elemental data mining problems
that intends to discover groups of items or values or patterns that co-occur
frequently in a dataset.

The term heart disease encompasses the diverse diseases that affect the
heart. Heart disease was then major cause of casualties in the United States,
England, Canada and Wales as in 2007. Heart disease kills one person every 34
seconds in the United States. Coronary heart disease, Cardiomyopathy and
Cardiovascular disease are some categories of heart diseases. The term

18
“cardiovascular disease” includes a wide range of conditions that affect the heart
and the blood vessels and the manner in which blood is pumped and circulated
through the body. Cardiovascular disease (CVD) results in severe illness,
disability, and death.

Predictive analytics include various statistical techniques from predictive

modeling, machine learning (ML) and data mining to make predictions based on
the current or historical data. The use of predictive analytics is in the customer
relationship management, healthcare industry and in many other fields. Deep
learning has a significant impact on the predictive analytics. There are many
models in the predictive modeling such as Naive bayes, Logistic regression,
Neural networks, Support Vector Machine, Classification and Regression trees
etc.

The risk factors for coronary heart disease (CHD) or atherosclerosis is

identified using the in-built imputation algorithm and particle swarm
optimization. It is obtained that physical inactivity also forms the risk factor of
CHD. The decision rules are used in predicting the risk factors of heart disease.
The risk assessment in patients suffering from congestive heart failure is done by
automatic classifier. Using the measure called Long-term heart rate variability,
their classifier classifies the lower risk patients from higher risk patients
automatically. Classification and Regression tree (CART) is useful in classifying
the higher risk patients and the lower risk patients. The heart rate variability is
also major important factor in finding the risk assessment of congestive heart
failure. The risk assessment of heart targeting the reduction of coronary heart
disease (CHD) events are classified into before the event and after the event. The
event before the CHD and after the event which are non-modifiable and the
modifiable are identified. The events are percutaneous coronary intervention
(PCI), myocardial infarction (MI) and coronary artery bypass graft surgery
(CABG). The C4.5 decision tree algorithm is used for these three events of

19
coronary heart disease. Association rules are used for prediction of heart disease.
These association rules are applied on the medical dataset and it generates many
rules which are irrelevant. In order to identify the rules that are truly essential for
predicting the heart disease are identified by using search constraints which
searches the association rules in training dataset and finally validates on the test
set. The hybrid system is used with the global optimization of genetic algorithm
and their system is used to initialize the neural network weights. A multi-layered
feed-forward network is used. The input nodes, hidden nodes and output nodes
are 12, 10 and 2 respectively. The input nodes are basically the risk factors which
are used in predicting the heart disease. The risk factors of heart attack otherwise
called myocardial infarction is identified based on the decision trees and Apriori
algorithm. Based on these methods, the risk factors which are identified as
efficient in the detection of heart attack are chest pain, diabetes, smoking, gender
and physical inactivity, age, lipids, cholesterol, triglyceride, blood pressure.

3.1.1 Disadvantages
● Low accuracy
● Highly expensive
● Sensitive to noise
● Highly complicated
3.2 PROPOSED SYSTEM

Now a day’s artificial neural network (ANN) has been widely used as a tool
for solving many decision modeling problems. A multilayer perception is a feed
forward ANN model that is used extensively for the solution of a no. of different
problems. An ANN is the simulation of the human brain. It is a supervised
learning technique used for non-linear classification coronary heart disease is
major epidemic in India and Andhra Pradesh is in risk of coronary heartdisease.
Clinical diagnosis is done mostly by doctor’s expertise and patients wereasked to
take no. of diagnosis tests. But all the tests will not contribute towards

20
effective diagnosis of disease. Feature subset selection is a pre-processing step
used to reduce dimensionality, remove irrelevant data. In this project we introduce
a classification approach which uses ANN and feature subset selection for the
classification of heart disease. PCA is used for pre-processing and to reduce no.
Of attributes which indirectly reduces the no. of diagnosis tests whichare needed
to be taken by a patient. We applied our approach on Andhra Pradeshheart disease
data base. Our experimental results show that accuracy improved over traditional
classification techniques. This system is feasible and faster and more accurate for
diagnosis of heart disease.

The proposed system is a classification approach which combines multi-

layer perception with back propagation learning algorithm and feature selection
for classification of heart disease with reduced no of attributes. Our approach
proves classification and determines the attributes which contribute more towards
the predication of heart disease which indirectly reduces no. of diagnosis test
which are needed tube taken by a patient.

3.3 ARTIFICIAL NEURAL NETWORK

An Artificial Neural Network also called as neural network is a

mathematical model based on biological neural networks. Artificial neural
network is based on observation of a human brain. Human brain is very
complicated web of neurons. Analogically artificial neural network is an
interconnected set of three simple units namely input, hidden and output unit. The
attributes that are passed as input to the next form a first layer. In medical
diagnosis patients risk factors are treated as input to the artificial neural network.

An artificial neural network (ANN) is the simulation of human brain and is

being applied to an increasingly number of real-world problems. Neural networks
as a tool we can mine knowledgeable data from data ware house. ANN are trained
to recognize, store and retrieve patterns to solve combinatorial

21
optimization problems. Pattern recognition and function Estimation abilities make
ANN prevalent utility in data mining. Their main advantage is that they cansolve
problems that are too complex for conventional technologies. Neural networks are
well suited to problem like pattern recognition and forecasting. ANN are used to
extract useful patterns from the data and infer rules from them.These are useful in
providing information on associations, classifications and clustering.

Fig 3.1: Example ANN

Popular neural network algorithms are Hopfield, Multilayer perception,

counter propagation networks, radial basis function and self-organizing maps etc.
The feed forward neural network was the first and simplest type of artificial neural
network consists of 3 units input layer, hidden layer and output layer. There are
no cycles or loops in this network. A neural network has to be configured to
produce the desired set of outputs. Basically, there are three learning situations
for neural network. 1) Supervised Learning, 2) Unsupervised Learning, 3)
Reinforcement learning the perception is the basic unit of an

22
artificial neural network used for classification where patterns are linearly
separable. Output node is used to represent the model output the nodes in neural
network architecture are commonly known as neurons. Each input node is
connected to output node via a weighted link. This is used to emulate the strength
of synaptic connection between neurons.

Step 1: Ret D= {{Xi, Yi}/i=1, 2, 3 n}} be the set of training example.

Step 2: Initialize the weight vector with random value.

Step 3: Repeat.

Step 4: For each training sample (Xi, Yi) ∈ D.

Step 5: Compute the predicted output Yi^ (k)

Step 6: For each weight we do.

Step 7: Update the weight we (k+1) = Wj(k) + (y i - yi^ (k))xij.

Step 8: End for.

Step 9: End for.

Step 10: Until stopping criteria is met

The main function of artificial neural network is prediction. The

effectiveness of artificial neural network was proven in medicine. The noteworthy
achievement of neural network was in the application of coronary heart disease

3.4 ADVANTAGES
● High Accuracy and Independent from prior assumptions about the
distribution of the data.
● Noise tolerance & Ease of Maintenance.
● ANN can be implemented in parallel hardware

23
 CHAPTER 4

SYSTEM DESIGN

4.1 SYSTEM ARCHITECTURE

Data Pre- Training Feature

Collection processing Data Selection

Model Building
Evaluation ANN

Test Data Prediction

Performance
Measures

Fig 4.1: System Architecture

4.2 HEART DISEASE PREDICTION
Data pre-processing evolves the step of the feature extraction and feature
selection. Feature selection means selection of the feature from the big set of
features to the small and relevant set of features, but feature extraction transforms
the big set of features to the new generated small set of features. Feature selection
gives original features, but feature extraction gives transformed features. We have
used feature extraction by one of the well-known methods, Principal component
analysis (PCA).
Principle component analysis (PCA) is used for transforming the big
dimension to lower one. For this Correlation Matrix Eigenvalues and Eigen
Vectors is calculated based on which new feature is extracted with more
relevancy. A perceptron model is made to elaborate working. It contains only a

24
single neuron with 13 inputs inside and a step function is used as a transfer
function. There is a bias also which controls the y-intercepts and to each input,
there is a separate weight. Finally, ANN is used for classification of heart disease.
4.3 MODULES
The modules involved in the proposed ANN based heart disease prediction
are explained as follows:
4.3.1 Data Collection
The data used in this project is obtained from the Cleveland Heart Disease
database. A total of 297 records with 14 medical attributes which is used to predict
the heart disease.
This database contains 76 attributes, but all published experiments refer to
using a subset of 14 of them. In particular, the Cleveland database is the only one
that has been used by ML researchers to this date. The "goal" field refers to the
presence of heart disease in the patient. It is integer valued from 0 (no presence)
to 4. Experiments with the Cleveland database have concentrated on simply
attempting to distinguish presence (values 1,2,3,4) from absence (value 0).
The names and social security numbers of the patients were recently
removed from the database, replaced with dummy values.
One file has been "processed", that one containing the Cleveland database.
All four unprocessed files also exist in this directory.
Only 14 attributes used:
1. #3 (age)
2. #4 (sex)
3. #9 (cp)
4. #10 (trestbps)
5. #12 (chol)
6. #16 (fbs)
7. #19 (restecg)
8. #32 (thalach)

25
 9. #38 (exang)
10. #40 (oldpeak)
11. #41 (slope)
12. #44 (ca)
13. #51 (thal)
14. #58 (num) (the predicted attribute)
The description of dataset is tabulated in Table 4.1.
Table 4.1: Description of Dataset

S.No Attribute Name Description

1 Age Age in years

2 Sex 1= male, 0= female

Chest pain type (1= typical angina, 2= atypical

3 Cp
angina, 3= non-anginal pain, 4= asymptomatic

Resting blood pressure (in mm Hg on admission

4 Trestbps
to hospital)

5 Chol Serum Cholesterol in mg/dl

Fasting blood sugar >120 mg/dl (1= true, 0=

6 Fbs
false)

Resting electrographic results (0= normal, 1=

7 Restecg having ST-T wave abnormality, 2= left
ventricular hypertrophy

8 Thalach Maximum heart rate

9 Exang Exercise induced angina

ST depression induced by exercise relative to

10 Oldpeak
Rest

Slope of the peak exercise ST segment (1=

11 Slope
upsloping, 2= flat, 3= downsloping)

12 Ca Number of blood vessels colored by fluoroscopy

26
 13 Thal 3= normal, 6= fixed defect, 7= reversible effect

14 Num Class (0= healthy, 1= have heart disease)

4.3.2 Training and Testing Dataset

The dataset of 297 records is divided into training and testing dataset. The
training dataset is used to build a predictive relationship and it is the set of
examples that is used for learning and to fit weights of the classifier. The test set
is set of examples which is used to evaluate the performance of a fully-specified
classifier. The training and testing dataset is divided into 75% and 25%
respectively.

Here, the model will be trained using the datasets and tested for finding the
accuracy of the model. Optimization will be done to improve the accuracy if
needed. In machine learning, a common task is the study and construction of
algorithms that can learn from and make predictions on data. Such algorithms
work by making data-driven predictions or decisions, through building a
mathematical model from input data. The data used to build the final model
usually comes from multiple datasets. In particular, three data sets are commonly
used in different stages of the creation of the model.

The model (e.g., a neural net classifier) is trained on the training dataset
using a supervised learning method (e.g., gradient descent or stochastic gradient
descent). In practice, the training dataset often consist of pairs of an input vector
(or scalar) and the corresponding output vector (or scalar), which is commonly
denoted as the target (or label). The current model is run with the training dataset
and produces a result, which is then compared with the target, for each input
vector in the training dataset. Based on the result of the comparison and the
specific learning algorithm being used, the parameters of the model are adjusted.

27
The model fitting can include both variable selection and parameter estimation.
Successively, the fitted model is used to predict the responses for the observations
in a second dataset called the validation dataset.

The validation dataset provides an unbiased evaluation of a model fit on the

training dataset while tuning the model's hyper parameters (e.g., the number of
hidden units in a neural network). Validation datasets can be used for
regularization by early stopping: stop training when the error on the validation
dataset increases, as this is a sign of over fitting to the training dataset. This simple
procedure is complicated in practice by the fact that the validation dataset's error
may fluctuate during training, producing multiple local minima. This
complication has led to the creation of many ad-hoc rules for deciding when over
fitting has truly begun. Finally, the test dataset is a dataset used to provide an
unbiased evaluation of a final model fit on the training dataset. If the data in the
test dataset has never been used in training (for example in cross-validation), the
test dataset is also called a holdout dataset.

4.3.3 Building Neural Network

The neural network is a computational model based on biological neural

networks. Artificial neural networks (ANN) are based on observation of a human
brain. Human brain is very complicated web of neurons. Analogically ANN is an
interconnected set of three units such as input, hidden and output unit. In medical
diagnosis, the patient's risk factors or attributes is used as an input. The
effectiveness of artificial neural network was proven in medicine. ANN are used
in predicting coronary heart disease. Here the input layer consisting of 8 neurons
corresponds to 8 significant attributes. There is one output class variable which
takes the value either 0 or 1. The value 0 represents that the individual is not
suffering from heart disease and the value 1 represents that the individual suffers
from heart disease. The number of nodes used in the hidden layer are 3.

28
 The Heart Disease Dataset consists of 14 input attributes including age,
gender, type of chest pain, blood pressure, cholesterol, blood sugar level,
electrocardiograph result, maximum heart rate, exercise induced angina, old peak,
Slope, number of vessels colored, thal. The dataset contains 303 instances, from
which 2 instances for a number of vessels colored attribute and 4 instances for
thal attribute are missing, which are filled by their mean value for the dataset
respectively. The prediction attribute consists of 2 classes ranging from integer
value 0 - 1 where 0 indicate no heart disease and the integer value from 1 indicate
the presence of heart disease. The feature selection from 14 input parameter by
backward elimination resulted in a total of 14 significant input parameters which
include gender, type of chest pain, blood pressure, blood sugar level,
electrocardiograph result, maximum heart rate, exercise induced angina, old peak,
Slope, number of vessels colored, thal.

4.3.4 Performance Measures

The performance measures of neural network are calculated using various

measures such as accuracy, specificity and sensitivity. The main advantage of
neural networks is high accuracy. The applications of neural network are
accounting, medicine, fraud detection etc. Based on the learned network or
training dataset, the neural network is able to predict the presence or absence of
heart disease for the testing dataset.

 Precision: It is the percentage of results that are relevant and is defined as:
𝑇𝑟𝑢𝑒𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒
𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 =
𝑇𝑟𝑢𝑒𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒 + 𝐹𝑎𝑙𝑠𝑒𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒

 Accuracy: Formally, accuracy has the following definition:

𝑇𝑟𝑢𝑒𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒 + 𝑇𝑟𝑢𝑒𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒
𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 =
𝑇𝑜𝑡𝑎𝑙

29
 Sensitivity: The sensitivity of a test is its ability to determine the patient cases
correctly. Mathematically, this can be stated as:
𝑇𝑟𝑢𝑒𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒
𝑆𝑒𝑛𝑠𝑖𝑡𝑖𝑣𝑖𝑡𝑦 =
𝑇𝑟𝑢𝑒𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒 + 𝐹𝑎𝑙𝑠𝑒𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒

 Specificity: The specificity of a test is its ability to determine the healthy cases
correctly. Mathematically, this can be stated as:

𝑇𝑟𝑢𝑒𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒
𝑆𝑝𝑒𝑐𝑖𝑓𝑖𝑐𝑖𝑡𝑦 =
𝑇𝑟𝑢𝑒𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒 + 𝐹𝑎𝑙𝑠𝑒𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒

30
 CHAPTER 5

SYSTEM DESCRIPTION

5.1 HARDWARE DESCRIPTION

System : Pentium IV 3.5 GHz or Latest Version.

Hard Disk : 40 GB.
Monitor : 14’ Color Monitor.
Mouse : Optical Mouse.
Ram : 1 GB.
5.2 SOFTWARE DESCRIPTION

Operating system : Windows 10

Coding Language : Python Programming
Tools used : Jupyter Notebook

5.2.1 Python Programming

Python is an interpreted, high-level and general-purpose programming

language. Created by Guido van Rossum and first released in 1991, Python's
design philosophy emphasizes code readability with its notable use of significant
whitespace. Its language constructs and object-oriented approach aim to help
programmers write clear, logical code for small and large-scale projects. Python
is dynamically typed and garbage-collected. It supports multiple programming
paradigms, including structured (particularly, procedural), object-oriented, and
functional programming. Python is often described as a "batteries included"
language due to its comprehensive standard library. Python interpreters are
available for many operating systems. A global community of programmers
develops and maintains CPython, a free and open-source reference
implementation. A non-profit organization, the Python Software Foundation,
manages and directs resources for Python and CPython development.

31
 Python is a multi-paradigm programming language. Object-oriented
programming and structured programming are fully supported, and many of its
features support functional programming and aspect-oriented programming
(including by metaprogramming and metaobjects (magic methods)). Many other
paradigms are supported via extensions, including design by contract and logic
programming.

Python uses dynamic typing and a combination of reference counting and

a cycle-detecting garbage collector for memory management. It also features
dynamic name resolution (late binding), which binds method and variable names
during program execution.

Python's design offers some support for functional programming in the Lisp
tradition. It has filter, map, and reduce functions; list comprehensions,
dictionaries, sets, and generator expressions. The standard library has two
modules (itertools and functools) that implement functional tools borrowed from
Haskell and Standard ML.

The language's core philosophy is summarized in the document The Zen of

Python (PEP 20), which includes aphorisms such as:

• Beautiful is better than ugly.

• Explicit is better than implicit.
• Simple is better than complex.
• Complex is better than complicated.
• Readability counts.

Rather than having all of its functionality built into its core, Python was
designed to be highly extensible. This compact modularity has made it
particularly popular as a means of adding programmable interfaces to existing
applications. Van Rossum's vision of a small core language with a large standard
library and easily extensible interpreter stemmed from his frustrations with ABC,
32
which espoused the opposite approach.

Python strives for a simpler, less-cluttered syntax and grammar while

giving developers a choice in their coding methodology. In contrast to Perl's
"there is more than one way to do it" motto, Python embraces a "there should be
one—and preferably only one—obvious way to do it" design philosophy. Alex
Martelli, a Fellow at the Python Software Foundation and Python book author,
writes that "To describe something as 'clever' is not considered a compliment in
the Python culture."

Python's developers strive to avoid premature optimization, and reject

patches to non-critical parts of the CPython reference implementation that would
offer marginal increases in speed at the cost of clarity. When speed is important,
a Python programmer can move time-critical functions to extension modules
written in languages such as C, or use PyPy, a just-in-time compiler. Cython is
also available, which translates a Python script into C and makes direct C-level
API calls into the Python interpreter.

An important goal of Python's developers is keeping it fun to use. This is

reflected in the language's name—a tribute to the British comedy group Monty
Python—and in occasionally playful approaches to tutorials and reference
materials, such as examples that refer to spam and eggs (from a famous Monty
Python sketch) instead of the standard foo and bar.
A common neologism in the Python community is pythonic, which can
have a wide range of meanings related to program style. To say that code is
pythonic is to say that it uses Python idioms well, that it is natural or shows
fluency in the language, that it conforms with Python's minimalist philosophy and
emphasis on readability. In contrast, code that is difficult to understand or reads
like a rough transcription from another programming language is called
unpythonic. Users and admirers of Python, especially those considered
knowledgeable or experienced, are often referred to as Pythonistas.
33
5.2.2 Jupyter Notebook

Project Jupyter is a project and community whose goal is to "develop open-

source software, open-standards, and services for interactive computing across
dozens of programming languages". It was spun off from IPython in 2014 by
Fernando Pérez. Project Jupyter's name is a reference to the three core
programming languages supported by Jupyter, which are Julia, Python and R, and
also a homage to Galileo's notebooks recording the discovery of the moons of
Jupiter. Project Jupyter has developed and supported the interactive computing
products Jupyter Notebook, JupyterHub, and JupyterLab.

Project Jupyter's operating philosophy is to support interactive data science

and scientific computing across all programming languages via the development
of open-source software. According to the Project Jupyter website, "Jupyter will
always be 100% open-source software, free for all to use and released under the
liberal terms of the modified BSD license".

Jupyter Notebook (formerly IPython Notebooks) is a web-based interactive

computational environment for creating Jupyter notebook documents. The
"notebook" term can colloquially make reference to many different entities,
mainly the Jupyter web application, Jupyter Python web server, or Jupyter
document format depending on context. A Jupyter Notebook document is a JSON
document, following a versioned schema, containing an ordered list of
input/output cells which can contain code, text (using Markdown), mathematics,
plots and rich media, usually ending with the ".ipynb" extension.

A Jupyter Notebook can be converted to a number of open standard output

formats (HTML, presentation slides, LaTeX, PDF, ReStructuredText, Markdown,
Python) through "Download As" in the web interface, via the nbconvert library or
"jupyter nbconvert" command line interface in a shell. To simplify visualisation
of Jupyter notebook documents on the web, the nbconvert library is provided as a
service through NbViewer[10] which can take a URL to any publicly available
34
notebook document, convert it to HTML on the fly and display it to the user.
Jupyter Notebook provides a browser-based REPL built upon a number of popular
open-source libraries:

● IPython
● ØMQ
● Tornado (web server)
● jQuery
● Bootstrap (front-end framework)
● MathJax

Jupyter Notebook can connect to many kernels to allow programming in

different languages. By default, Jupyter Notebook ships with the IPython kernel.
As of the 2.3 release (October 2014), there are currently 49 Jupyter-compatible
kernels for many programming languages, including Python, R, Julia and Haskell.

The Notebook interface was added to IPython in the 0.12 release, renamed
to Jupyter notebook in 2015 (IPython 4.0 – Jupyter 1.0). Jupyter Notebook is
similar to the notebook interface of other programs suchas Maple, Mathematica,
and SageMath, a computational interface style that originated with Mathematica
in the 1980s.Colaboratory (also known as Colab) is a free Jupyter notebook
environment that runs in the cloud and stores its notebooks on Google Drive.
Colab was originally an internal Google project; an attempt was made to open
source all thecode and work more directly upstream, leading to the developmentof
the "Open in Colab" Google Chrome extension, but this eventually ended, and
Colabdevelopment continued internally. As of October 2019, the Colaboratory UI
only allows for the creation of notebooks with Python 2 and Python 3
kernels;however, an existing notebook whose kernelspec is IR or Swift will also
work, since both R and Swift are installed in the container. Julia language can also
workon Colab (with e.g., Python and GPUs; Google's tensor processing units also
workwith Juliaon Colab).

35
 CHAPTER 6

SYSTEM IMPLEMENTATION

6.1 SOFTWARE IMPLEMENTATION

Implementation includes all those activities that take place to convert from
the old system to the new. The old system consists of manual operations, which
is operated in a very different manner from the proposed new system. A proper
implementation is essential to provide a reliable system to meet the requirements
of the organizations. An improper installation may affect the success of the
computerized system. The implementation steps are listed below.

36
6.1.1 Implementation Methods

There are several methods for handling the implementation and the
consequent conversion from the old to the new computerized system.

The most secure method for conversion from the old system to the new
system is to run the old and new system in parallel. In this approach, a person may
operate in the manual older processing system as well as start operating the new
computerized system. This method offers high security, because even if there is a
flaw in the computerized system, we can depend upon the manual system.
However, the cost for maintaining two systems in parallel is very high. This
outweighs its benefits.

Another commonly method is a direct cut over from the existing manual
system to the computerized system. The change may be within a week or within
a day. There are no parallel activities. However, there is no remedy in case of a
problem. This strategy requires careful planning.

A working version of the system can also be implemented in one part of

the organization and the personnel will be piloting the system and changes can be
made as and when required. But this method is less preferable due to the loss of
entirety of the system.

The implementation plan includes a description of all the activities that must occur
to implement the new system and to put it into operation. It identifies the personnel
responsible for the activities and prepares a time chart for implementing the
system.
The implementation plan consists of the following steps.

● List all files required for implementation.

37
 ● Identify all data required to build new files during the implementation.
● List all new documents and procedures that go into the new system.
The implementation plan should anticipate possible problems and must be
able to deal with them. The usual problems may be missing documents; mixed
data formats between current and files, errors in data translation, missing data etc.

For heart disease prediction, it involves gathering comprehensive datasets

that include patient information and relevant clinical features. These datasets can
be sourced from medical records, health surveys, hospital databases, and public
health repositories. Key data points might include demographic information (age,
gender, ethnicity), medical history (previous heart conditions, family history of
heart disease, diabetes, hypertension), lifestyle factors (smoking status, alcohol
consumption, physical activity levels, diet), and clinical measurements (blood
pressure, cholesterol levels, heart rate, ECG results, and other diagnostic test
results). Ensuring the data is accurate, complete, and representative of the target
population is critical. Data collection may involve obtaining necessary
permissions and ensuring compliance with data privacy regulations such as
HIPAA or GDPR.

6.1.2 Import Libraries

Pandas is for reading our dataset, NumPy is for working with arrays and
also performing linear algebra calculations, seaborn for visualizations, and also
Matplotlib is for visualization, and lastly, warning messages are typically issued
in situations where it is useful to alert the user of some condition in a program,
where that condition (normally) doesn’t warrant raising an exception and
terminating the program.
So, our aim here is to analyze our data and report our findings through
visualizations and the code below allows us to check if we have any missing
values in our dataset before going further with the analysis.
38
6.1.3 Analyze and Visualize Pandas Data Structure

The describe() function computes a summary of statistics pertaining to the Data

Frame columns. This function gives the mean, std, and IQR values.

Data Cleaning: Identifying and correcting errors or inconsistencies in the data.

This includes handling missing values, removing duplicates, and correcting
erroneous entries.

Normalization/Standardization: Scaling the data so that features with different units

or magnitudes do not disproportionately influence the model. Common techniques
include min-max normalization and z-score standardization.

Encoding Categorical Variables: Converting categorical data (e.g., gender,

smoking status) into numerical format using methods like one-hot encoding or
label encoding.

Splitting the Data: Dividing the dataset into training, validation, and test sets.
Typically, the data is split in a ratio such as 70-20-10, where 70% is used for
training, 20% for validation, and 10% for testing.

The data preprocessing involves several key steps. First, the describe() function
is used to generate a statistical summary, including mean and standard deviation.
Next, data cleaning is performed to correct errors, handle missing values, and
remove duplicates. Then, normalization or standardization is applied to ensure
that features are on a comparable scale. Categorical variables are encoded into
numerical formats using techniques like one-hot or label encoding. Finally, the
dataset is split into training, validation, and test sets, typically in a 70-20-10 ratio,
for effective model development and evaluation.
39
6.1.4 Exploratory Data Analysis

The unique() function on the output variable shows the data type and
present unique values. Here we can notice that the output variable is imbalanced
with a small fraction which does make any difference in the development of the
algorithm.
6.1.5 Checking Relationship among Variables

Now checking the relationship between variables using the seaborn relplot
chart and from the result, we can deduct that people between the age of 50 years
and 65 are experiencing cp = 0 (AGINA) which is the most common pain among
the list. Also, we can conclude that these heart disease and cholesterol levels
indicate that it is mostly among women than men.
6.1.6 Checking and Dropping Duplicates

The number of duplicated entries is 1 row and 14 columns, which basically

means that we only have one duplicated entry hence we drop.
6.1.7 Outlier Detection and Removal

Here we used the z-score to detect outliers in our dataset and I used a
threshold of the 3 to filter the outliers. our dataset of 303 entries was reduced to
287 entries
6.1.8 Pre-processing

We are preprocessing our data before building our model and we are going
to separate them into three categories namely:
6.1.8.1 Split the data into training and testing sets
6.1.8.2 Scaling the features into an unformal range.
6.1.8.3 Scaling our data

40
Train test split
Here at this stage, we enter another phase into the development of our
machine learning model which is training and testing, we split the data into two
sections; training and testing. we do this because we can’t use the data to train
and test our model for too, hence we separate data for training and the other for
testing with our model.
Scaling
We scale the data so that when feeding the data into the machine learning
model it will be uniform which also positively impacts our model accuracy.

Building an ANN involves several key steps:

Designing the Network Architecture: Determining the structure of the ANN,

including the number of layers (input, hidden, and output layers), and the number
of neurons in each layer. A typical architecture for heart disease prediction might
include:

o Input Layer: Corresponds to the number of features selected.

o Hidden Layers: One or more layers with a specified number of neurons,

using activation functions such as ReLU (Rectified Linear Unit).

o Output Layer: A single neuron with a sigmoid activation function for

binary classification (presence or absence of heart disease).

Initializing Weights: Setting initial weights for the network, often using
methods like Xavier initialization to facilitate efficient training.

Forward Propagation: Passing the input data through the network to obtain
predictions. Each neuron calculates a weighted sum of its inputs and applies an
activation function.

Backpropagation: Adjusting the weights based on the error between

predicted and actual outcomes. This process involves calculating gradients and
41
updating weights using optimization algorithms such as stochastic gradient descent
(SGD).

Training the Network: Iteratively feeding the training data through the
network and updating weights until the model converges to a minimum error.

6.1.9 Fitting an ANN model

ANN is an algorithm that can both perform classification task. A linear

model is used to predict categorical variables with the help of dependent variables,
consider you need to determine if a variable belongs to any of the two classes
available. The algorithm checks relationship and probability of the dependent
variables in consideration of the classes available.

6.1.10 Confusion Matrix

Confusion Matrix is the visual representation of the Actual VS Predicted

values. It measures the performance of our Machine Learning classification
model and looks like a table-like structure.

6.1.11 ANN Model Classification and Saving

Now the developed ANN predictive heart disease model is evaluated by

classifying a test data and if the prediction result is accurate as desired, then the
developed ANN model can be saved for future prediction problems.

42
 CHAPTER 7

SYSTEM TESTING

System testing is a critical aspect of Software Quality Assurance and

represents the ultimate review of specification, design and coding. Testing is a
process of executing a program with the intent of finding an error. A good test is
one that has a probability of finding an as yet undiscovered error. The purpose of
testing is to identify and correct bugs in the developed system. Nothing is
complete without testing. Testing is the vital to the success of the system.
In the proposed system of heart disease prediction, the code testing the
logic is tested using the developed ANN model which has been running on Python
programming language in Jupyter Notebook. For this, every module of the
program is executed to find an error. To perform specification test, the
examination of the specifications stating what the program should do and how it
should perform under various conditions. Unit testing focuses first on the modules
in the proposed system to locate errors. This enables to detect errors in the coding
and logic that are contained within that module alone. Those resultingfrom the
interaction between modules are initially avoided. In unit testing step each module has
to be checked separately. Hence, in this project, the modules such as the dataset
preparation, training, testing, model building and performanceevaluation are
tested one by one.
System testing does not test the software as a whole, but rather than
integration of each module in the system. The primary concern is the
compatibility of individual modules. One has to find areas where modules have
been designed with different specifications of data lengths, type and data element
name. Testing and validation are the most important steps after the
implementation of the heart disease prediction system. The system testing is
performed to ensure that there are no errors in the implemented proposed heart

43
disease prediction system. The software must be executed several times in order
to find out the errors in the different modules of the proposed system.
Validation refers to the process of using the new software for the developed
system in a live environment i.e., the jupyter notebook file of the proposed heart
disease prediction model, in order to find out the errors in it. The validation phase
reveals the failures and the bugs in the heart disease prediction system. It will
become to know about the practical difficulties, the heart disease prediction model
faces when operated in the jupyter notebook. By testing the code of the
implemented software, the logic of the program can be examined. A specification
test is conducted to check whether the specifications like the hardware
specifications like system, monitor, RAM, ROM, mouse, keyboard, etc. and
software specifications like python 3.7 and jypyter notebook, stating the program
are performing under various conditions. Apart from these tests, there are some
special tests conducted which are given below:
Peak Load Tests: This determines whether the proposed heart disease
prediction system will handle the volume of activities when the system is at the
peak of its processing demand. The test has revealed that the new software for the
agency is capable of handling the demands at the peak time.
Storage Testing: This determines the capacity of the proposed heart
disease prediction system to store transaction data on a disk or on other files. The
proposed software has the required storage space available, because of the use of
a number of hard disks.
Performance Time Testing: This test determines the length of the time
used by the heart disease prediction to process transaction data.
In this phase the heart disease prediction model developed testing is
exercising the software to uncover errors and ensure the system meets defined
requirements. Testing may be done at 4 levels
• Unit Level

44
• Module Level
• Integration & System
• Regression

7.1 UNIT TESTING

A Unit corresponds to a screen/form in the package. Unit testing focuses on
verification of the predictive model classes like heart disease positive or negative.
This testing includes testing of control paths, interfaces, local data structures,
logical decisions, boundary conditions, and error handling. Unit testing use Test
Drivers, which are control programs to co-ordinate heart image inputs and
outputs, and Test stubs like testing module, which replace low-level modules. A
stub is a dummy subprogram.
7.2 VALIDATION TESTING
In this requirement established as part of software requirements analysis
are validated against the software that has been constructed. Validation testing
provides final assurance that the heart disease model meets all functional,
behavioral and performance requirements. Validation can be defined in many
ways but a simple definition is that validation succeeds when software Function
in a manner that can be reasonably by the customer, i.e., the proposed heart
disease predictive model accurately predicts the heart disease availability in the
input test image.

1. Validation test criteria

2. Configuration review

3. Alpha and Beta testing (conducted by end user)

7.3 MODULE LEVEL TESTING

Module Testing is done using the test cases prepared earlier. Module is
defined during the time of designing the heart disease predictive model in python
environment.

45
7.4 INTEGRATION & SYSTEM TESTING
Integration testing is used to verify the combining of the software modules
like dataset preparation, training, testing, model building and performance
evaluation. Integration testing addresses the issues associated with the dual
problems of verification and program construction. System testing is used to
verify, whether the predictive model meets the requirements. System testing is
actually a series different test whose primary purpose is to full exercise the
computer base system, where the software and other system elements are tested
as whole. To test computer software, we spiral out along streamlines that broadens
the scope of testing with each turn.

The last higher-order testing step falls outside the boundary of software
engineering and in to the broader context of computer system engineering. The
proposed predictive model of heart disease availability, once validated, must be
combining with order system elements (e.g., hardware, people, databases). System
testing verifies that all the elements mesh properly and that overall system
function/performance is achieved.
1. Recovery Testing
2. Security Testing
3. Stress Testing
7.5 REGRESSION TESTING
Each modification in software impacts unmodified areas, which results
serious injuries to that software. So, the process of re-testing for rectification of
errors due to modification is known as regression testing.
Installation and Delivery: Installation and Delivery is the process of delivering
the proposed predictive model and tested model to the customer.
Acceptance and Project Closure: Acceptance is the part of the project by which
the customer accepts the product. This will be done as per the Project Closure,
once the customer accepts the developed heart disease prediction model, closure
of the project is started. This includes metrics collection, PCD, etc.

46
 CHAPTER 8

RESULTS AND DISCUSSION

8.1 MODULES IMPORT

The necessary modules are imported and n sample data is retrieved and shown
from the data set.
8.2 DATASET VISUALIZATION

The description of the attributes in the data set are shown.

47
Preprocessing is done by dropping the duplicates

A histogram of the values of the attributes are shown.

48
8.2 CORRELATION MATRIX

A plot for the target is depicted.

A correlation matrix is visualized.

49
8.3 FEATURE SELECTION

Application of feature selection algorithm Chi square.

8.4 ANN MODEL TRAINING

Training the model - Artificial Neural network.

50
 Output prediction after model training.

8.5 PREDICTION

Output prediction after model training.

51
8.6 CLASSIFICATION

Classification report for the trained model is depicted.

8.7 PERFORMANCE EVALUATION

Plot for Loss.

52
Plot for accuracy.

Roc curve plot.

53
Acquiring the required op through custom input.

54
 CHAPTER 9

CONCLUSION

In our project we have proposed a new feature selection method for heart
disease classification using ANN and various feature selection methods for
Andhra Pradesh Population. We applied different feature selection methods to
rank the attributes which contribute more towards classification of heart disease,
which indirectly reduces the no. of diagnosis tests to be taken by a patient. Our
experimental results indicate that on an average with ANN and feature subset
selection provides on the average better classification accuracy and
dimensionality reduction. Our proposed method eliminates useless and distortive
data. This research will contribute reliable and faster automatic heart disease
diagnosis system, where easy diagnosis of heart disease will save lives. Coronary
heart disease can be handled successfully if more research is encouraged in this
area.

55
REFERENCES

[1] Ahmed, Rizgar Maghdid, and Omar Qusay Alshebly. "Prediction and factors
affecting of chronic kidney disease diagnosis using artificial neural networks
model and logistic regression model." Iraqi Journal of Statistical Sciences
16.28 (2019): 140-159.
[2] Burse, Kavita, et al. "Various pre-processing methods for neural network
based heart disease prediction." Smart innovations in communication and
computational sciences. Springer, Singapore, 2019. 55-65.
[3] Ali, Farman, et al. "A smart healthcare monitoring system for heart disease
prediction based on ensemble deep learning and feature fusion." Information
Fusion 63 (2020): 208-222.
[4] Shahid, Nida, Tim Rappon, and Whitney Berta. "Applications of artificial
neural networks in health care organizational decision-making: A scoping
review." PloS one 14.2 (2019): e0212356.
[5] Khourdifi, Youness, and Mohamed Bahaj. "Heart disease prediction and
classification using machine learning algorithms optimized by particle
swarm optimization and ant colony optimization." Int. J. Intell. Eng. Syst.
12.1 (2019): 242-252.
[6] Khalil, Ahmed J., et al. "Energy Efficiency Predicting using Artificial Neural
Network." (2019).
[7] Ali, Liaqat, et al. "An Automated Diagnostic System for Heart Disease
Prediction Based on Chi-Square Statistical Model and Optimally Configured
Deep Neural Network." IEEE Access 7 (2019): 34938-34945.
[8] Mohan, Senthilkumar, Chandrasegar Thirumalai, and Gautam Srivastava.
"Effective heart disease prediction using hybrid machine learning
techniques." IEEE Access 7 (2019): 81542-81554.
[9] Latha, C. Beulah Christalin, and S. Carolin Jeeva. "Improving the accuracy
of prediction of heart disease risk based on ensemble classification
techniques." Informatics in Medicine Unlocked 16 (2019): 100203.
[10] Darmawahyuni, Annisa, Siti Nurmaini, and Firdaus Firdaus. "Coronary
Heart Disease Interpretation Based on Deep Neural Network." Computer
Engineering and Applications Journal 8.1 (2019): 1-12.
[11] Aqueel Ahmed, Shaikh Abdul Hannan, “Data Mining Techniques to Find
Out Heart Diseases: An Overview”, IJITEE.

56
[12] Monika Gandhi, Dr. Shailendra Narayan Singh, “Predictions in Heart
Disease Using Techniques of Data Mining”, IEEE.(2019)
[13] Marjia Sultana, Afrin Haider and Mohammad ShorifUddin, “Analysis of
Data Mining Techniques for Heart Disease Prediction”.
[14] M.A.Nishara Banu , B Gomathy, “Disease Predicting System Using Data
Mining Techniques” IJTRA.
[15] K Raj Mohan ,Ilango Paramasivam ,SubhashiniSathyaNarayan, “Prediction
and Diagnosis of Cardiovascular Disease – A Critical Survey”, IEEE (2019)
[16] Mohamed Elnoamany, Ashraf Dawood, Nahed Mohamed Momtaz, Waleed
Abdou, “Thrombospondin-1 Levels in Patients with Coronary Heart
Disease” WJCD (2021)
[17] Abhishek Rairikar, Vedant Kulkarni, Vikas Sabale , “Heart Disease
Prediction Using Data Mining Techniques”, IEEE (2018)
[18] Pahulpreet Singh Kohli, Shriya Arora, “Application of Machine Learning in
Disease Prediction”, IEEE (2019)
[19] Jyoti Soni, Ujma Ansari, “Predictive Data Mining for Medical Diagnosis:
An Overview of Heart Disease Prediction”, IJCA.
[20] Mudasir M Kirmani, “Cardiovascular Disease Prediction Using Data Mining
Techniques: A Review”, Oriental Scientific Publishing Co. (2019)
[21] Bin Deng, Min Guo "Risk Factors and Intervention Status of Cardiovascular
Disease in Elderly Patients with Coronary Heart Disease", Health( 2020)
[22] T. Velmurugan, U. Latha "Classifying Heart Disease in Medical Data Using
Deep Learning Methods", Journal of Computer and Communications (2021)
[23] Karaskov Alexander "Impact of chemical elements on heart failure
progression in coronary heart disease patients", SciRes.
[24] Jyoti Soni,Ujma Ansari "Predictive Data Mining for Medical Diagnosis: An
Overview of Heart Disease Prediction", International Journal of Computer
Applications((2019)
[25] Ammar Aldallal; Amina Abdul Aziz Al-Moosa "Using Data Mining
Techniques to Predict Diabetes and Heart Diseases", IEEE (2018)
[26] Monika Gandhi, Shailendra Narayan Singh "Predictions in Heart Disease
Using Techniques of Data Mining", IEEE.
[27] Durairaj M, Revathi V, "Prediction Of Heart Disease Using
Backpropagation MLP Algorithm", IJSTR (2019)

57
[28] M. Akhiljabbar,B.L.Deekshatulu " Classification of Heart Disease Using K-
Nearest Neighbor and Genetic Algorithm "Procedia Technology.
[29] ResulDas, IbrahimTurkoglu,"Effective diagnosis of heart disease through
neural networks ensembles", Expert Systems with Applications.
[30] Samir B Patel,Samir Tejalupadhyay,"Heart Disease Prediction Using
Machine learning and Data Mining Technique”, IJCSC(2019)
[31] Saima Safdar, Saad Zafar, “Machine Learning Based Decision Support
Systems (DSS) For Heart Disease Diagnosis: A Review”, SpringerLink.
[32] P.Santhi a, R.Ajayb"A Survey on Heart Attack Prediction Using Machine
Learning" Turkish Journal of Computer and Mathematics Education (2021).

58