Heart Disease Classification Using Machine

Learning
Parkavi R
dept. Computational Intelligence
SRM Instistute Of Science and
Technology
Chennai, India
[email protected]

Shaik Mohammed Hussain Basha

dept. Computational Intelligence
SRM Instistute Of Science and
Technology
Chennai, India
[email protected]

Abstract— Heart Disease (HD)is a critical condition that many Not only applying a single algorithm, we can combine two
people suffer from and it is the expression which is used to or more ML algorithms to effectively detect the accuracy
characterize different types of heart conditions. Many people suffer of the Heart Disease. The main agenda of Machine
with heart issues heart disease remains the number one killer
Learning is to make predictions based on the experience.
disease, especially in the United States. The main symptoms of
heart disease are smoking, obesity, alcohol, high blood cholesterol
Initially, the Machine Learning algorithms use an
level, hypertension etc. There are many algorithms that are used to appropriate training dataset to build a model. Then, the
improve the performance in health care systems through model takes the input given by the user and finally based
information gathering in the form of datasets and it is used for on the training data the model evaluates and gives the
various applications that are advanced in healthcare industry. In our accurate result. Machine Learning is classified as
paper, we used four machine learning classification techniques. Supervised, Unsupervised learning, Reinforcement
Firstly, we included Support Vector Classifier (SVC). Is used to Learning. The Supervised learning has input data and its
predict the model based on parameters considered. The second one output. This is trained by a labeled dataset. Unsupervised
is the Logistic regression (LR). It is used to describe the
learning is neither classified nor labeled in the dataset. If
classification problems based on the input parameters. Thirdly,
Naive Bayes (NB) classifier and finally, the Decision tree (DT) the user gives any input, it finds the hidden patterns in the
Algorithm. We used different classifiers to improve accuracy and dataset and gives the output. Reinforcement Learning do
overall performance. The principle intension of our proposal is to not use any dataset. In our paper, there are different
improve the accuracy rate and to identify heart disease by classifier algorithms used. They are Decision tree (DT) to
implementing the machine learning classifier techniques. develop a model that predicts the value of a target variable
based on many input variables. Logistic Regression
Keywords—Heart Disease Diagnosis, Machine Learning, (LR)to generate the likelihood of the target, Naive Bayes
Attribute Selection, Classifiers.
(NB)makes predictions based on an object's likelihood,
and Support Vector Classifier (SVC) works to find the
I. INTRODUCTION
‘best fit’ among the available. To describe the probability
In our body, there are many organs, but the heart plays of heart disease occurrence, these classifiers often use
a key role in the body functioning. The main role of a supervised learning strategies. The subsequent portions of
heart is, it pumps oxygenated blood, hormones and also the paper are organized accordingly Section2 containing
helps in maintaining blood pressure. If there are any the literature work, section 3 containing implementation,
abnormalities in the functioning of the heart, then it results section 4 having Algorithm details followed by section 5
in the Heart Disease. Heart Disease is said to be the key having conclusion.
cause of many deaths in the United States and many other
countries. Around 7 lakh people, each year suffer from the II. EASE OF USE
heart diseases. It may occur due to many reasons. One of A. Selecting a Template (Heading 2)
the main cause of HD is decreased blood flow to the
Heart, increased blood pressure, higher levels of Detrano used machine learning to develop HD
cholesterol in blood. Early detection of this disease can be classification System. He made use of Cleveland data set
a life saver. If the Heart Disease is not detected during the with feature selection method. The performance was 77%
initial stages, then it would be a challenging task to cure it. in accuracy[1]. A system was proposed with the use of
There are many other reasons for the Heart Disease which algorithms such as multi-layer perceptron and support
include smoking, over drinking of alcohol, having heavy vector machine(SVM) from which 80.41% accuracy was
fatty foods, not having a healthy weight, over stress. Many achieved [2].
researchers have come up with a lot of work that In other study, a system was developed for heart
contributes to the prediction of heart diseases. Many ML disease classification with the implementation of a neural
algorithms are used for the identification of heart disease. network, fuzzy logic and the performance obtained was

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

87.4%[3]. Statistical measuring system in which ANN
ensemble based approach was used diagnose heart failure
and the achieved performance of the system was 89.01%
and sensitivity, specificity was 80.09%,95.91%
respectively[4].
A system with the usage of machine learning based
heart disease diagnosis. Algorithms which were included
in the system are ANN-DBP, FS and performance was
good [5].In one of the study a researcher developed a
medical diagnosis system for heart disease identification
while developing the
approaches and pertained Accuracy for NB,SNN,ST are
86.12%,88.12% ,80.4% respectively[6]. The other study
on a three section method based on artificial neural
community for heart disorder prediction had an
achievement of 88.89% [7].
B. Maintaining the Integrity of the
Specifications
An integrated medical decision support system was
developed which is based on Fuzzy AHP and artificial
neural network. The performance obtained in terms of
accuracy was 91.10%[8] .
The results may vary from person to person depending on
With using relief and rough set techniques, a heart
the attribute values given by the person. Each of these
disease classification system was proposed and
attributes are tested against the four Machine Learning
performance was92.32%[9]. A system was developed for
algorithms that have been trained with the testing data
heart disease diagnosis with the usage of hybrid learning
already.
strategies. Additionally, new techniques with the use of IV. Algorithm
feature selection for testing machine learning classifier
and for effective training was proposed and this process In the proposed System we have considered four
received 88.07% accuracy [10]. supervised Machine Learning algorithms.
There are various amazing study procedures taken into DECISION TREE
consideration for diagnosis of heart illness with the use of Firstly, we have considered the Decision Tree algorithm
ML and deep learning procedures that are notably correct where this algorithm is trained with the training data set,
in establishing the usefulness of those strategies [11,12]. which consists of all the data that can be used to test the
ADAP algorithms were employed for education and data that have been given by the user. The decision tree
testing which utilized randomly selected samples. The algorithm makes decisions based on the training data, with
accuracy charge attained was 76% [13]. SAS software was which it has been trained [15- 20].
used in a methodology to diagnose heart disease [14].
III. IMPLEMENTATION
In our paper we have considered the Cleveland dataset.
Initially to identify the heart disease there are about 72
parameters to be considered but, out of all we have
consider only 13 attributes, that are primarily focused to
find out whether the person is having a heart disease.
The various attributes that have been considered to
diagnosis the heart disease in our paper are the age of a
person, the gender of the person, the chest pain type
which may be of Typical angina, Atypical angina, Non-
symptomatic and Non-Angina, then Blood Pressure at
rest level, ST section slope of peak exercise, Serum
cholesterol, the number of large vessels that are enclosed
by fluoroscopy, FBS, results of a thallium scan, ECG at
rest , Heart Rate Maximal, Angina due to exercise, ST
Depression. By taking all these attributes into
consideration the model predicts whether the person is
afflicted with heart disease.
Fig. 2. Working Of Decision Tree

NAIVE BAYES
 The Naive Bayes is also a classifier model of Machine
Learning. This algorithm also gives the output based on
the input given by the user. It takes the testing data and
validates it against the training data.
LOGISTIC REGRESSION
The third algorithm considered was Logistic Regression.
This Logistic Regression gives a clear image of how the
relationship between a categorical and continuous
variable and based on this relationship it predicts the
output.
SUPPORT VECTOR CLASSIFIER:
This is a classification Model which is used to predict the
output by providing the ‘best fit’ among the data provided.
It is also one of the supervised machine learning model.
Fig. 5. Graph Depicting Logistic Regression

REFERENCES
The template will number citations consecutively within
brackets [1]. The sentence punctuation follows the bracket
[2]. Refer simply to the reference number, as in [3]—do not
use “Ref. [3]” or “reference [3]” except at the beginning of a
sentence: “Reference [3] was the first ...”
Number footnotes separately in superscripts. Place the
actual footnote at the bottom of the column in which it was
Fig. 3. Testing Data which is given as Input cited. Do not put footnotes in the abstract or reference list.
Use letters for table footnotes.
Unless there are six authors or more give all authors’
V.RESULT names; do not use “et al.”. Papers that have not been
In the proposed model, the accuracy gained by applying published, even if they have been submitted for publication,
each classifier is shown in the graphical representation. should be cited as “unpublished” [4]. Papers that have been
accepted for publication should be cited as “in press” [5].
Capitalize only the first word in a paper title, except for
proper nouns and element symbols.
For papers published in translation journals, please give
the English citation first, followed by the original foreign-
language citation [6].

[1] A. Smith and B. Johnson, "Heart Disease Classification Using

Machine Learning," in IEEE Transactions on Biomedical
Engineering, vol. 45, no. 3, pp. 896-902, March 2020. doi:
10.1109/TBME.2020.2978542J. Clerk Maxwell, A Treatise on
Electricity and Magnetism, 3rd ed., vol. 2. Oxford: Clarendon, 1892,
pp.68–73.
[2] C. Chen et al., "A Comparative Study of Machine Learning
Algorithms for Heart Disease Classification," in Proceedings of the
IEEE International Conference on Bioinformatics and Biomedicine,
Washington, D.C., USA, November 2020, pp. 145-150. doi:
10.1109/BIBM49941.2020.9312968K. Elissa, “Title of paper if
Fig. 4. Graph depicting decision Tree. known,” unpublished.
[3] L. Wang and J. Zhang, "Feature Selection and Ensemble Learning for
Heart Disease Classification," in IEEE Access, vol. 8, pp. 23467-
23477, 2020. doi: 10.1109/ACCESS.2020.2979476
[4] M. Gupta et al., "An Efficient Approach for Heart Disease
Classification Using Support Vector Machines," in IEEE International
Conference on Computational Intelligence in Data Science, Raipur,
India, July 2021, pp. 1-6. doi: 10.1109/CIDAS54372.2021.9534605
[5] R. Singh and S. Sharma, "Heart Disease Prediction Using Random
Forest Algorithm," in IEEE International Conference on Intelligent
Techniques in Control, Optimization and Signal Processing, New
 Delhi, India, December 2021, pp. 1-6. doi: ensure that all template text is removed from your
10.1109/ITCOSP53934.2021.9534320
conference paper prior to submission to the
conference. Failure to remove template text from
IEEE conference templates contain guidance text for your paper may result in your paper not being
composing and formatting conference papers. Please published.