Heart Disease Prediction using Machine Learning

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 04 | Apr 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1283
Heart Disease Prediction using Machine Learning
Prathamesh Keni [1], Pratik Poshe [2], Kaustubh Latake [3], Prof. Rovina Dbritto [4]
[1] [2] [3] Student, Department of Information Technology, Universal college of Engineering, Vasai
[4] Professor, Department of Information Technology, Universal college of Engineering, Vasai
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract – Recently, the cases of heart disease have been
increasing rapidly, and it's essential to predict these diseases.
The detection of the conditions is difficult to perform, and it
should be performed precisely due to its sensitive matter. We
have developed an application to predict a person's
probability of getting heart disease. We have tested multiple
algorithms such as KNN, Support Vector machines, Naive
Bayes, Decision tree Classifier, and Random Forest to find the
probability. Logistic Regression showed good accuracy in
comparison to the other algorithms. The heart disease
prediction system enhances the medical care systems and
reduces pressure on them. The project is implemented on the
android app, and the backend is implemented in python.
Key Words: Algorithm, KNN, Logistic Regression, Vector
Machines, Naive Bayes, Decision treeClassifier,Random
Forest
1.INTRODUCTION
Machine learning enables software programmes to grow
increasingly effective at predicting outcomes without
explicitly programming them. Machine learning algorithms
estimate new output values by using past data as input [1].
Machine Learning is a diversefield,anditsimplementationis
increasing daily. Machine Learning includes multiple
classifiers, categorised as Supervised, Unsupervised and
Ensemble Learning which are used to predict and find the
precision of a given dataset. We will use the knowledge in
our project on Heart Disease Prediction (HDP) as it will
benefit many people.
Cardiovascular diseases (CVDs) are the primary cause of
mortality worldwide. In 2019, an estimated 17.9 million
individuals died from CVDs, accounting for 32% of global
fatalities. A heart attack or a stroke caused Eighty-five per
cent of these fatalities. [2].
The project's objective is to check whether the patient is
likely to have cardiovascular disease or not based on
attributes such as age, gender, height, weight, blood
pressure, glucose level, cholesterol level, etc. We selected a
dataset from Kaggle with the patient's medical history and
these attributes. Using this dataset, we can determine
whether the patient can have cardiovascular disease or not.
We trained the dataset using Logistic Regression, Random
Forest and Naive Bayes algorithms. The most efficient of
these algorithms was Logistic Regression, with an accuracy
of 72.85%. Once the model is trained, the classification of a
patient is very cost-efficient.
2. DATA SOURCE
We chose a data source that contains the medical history of
70000 patients. This dataset gave us much-needed
information such as age, gender, blood pressure, glucose
level, etc. and whether the patient has heart disease or not.
This data helps us in creating a machine learning model to
classify the patients. The dataset is first split into two parts,
Training and Testing. The training part is used to train the
model, and the testing part is used to test the trained model.
The dataset is split into 80:20, meaning that 80% of data is
used for training while 20% is used for testing. All attributes
in the dataset are listed below in 'Table 1'.
Table -1: Dataset Attributes
Feature Variable Type Variable Value Type
Age Objective Feature age int(days)
Height Objective Feature height int (cm)
Weight Objective Feature weight float(kg)
Gender Objective Feature gender Categorical
Code
Systolic
blood
pressure
Examination
Feature
ap_hi int
Diastolic
blood
pressure
Examination
Feature
ap_lo int
Cholester
ol
Examination
Feature
cholesterol Categorical
Code
Glucose Examination
Feature
gluc Categorical
Code
Smoking Subjective
Feature
smoke binary
Alcohol
intake
Subjective
Feature
Physical Subjective
Feature
active binary
Presence
or
absence of
cardiovasc
ular
disease
Target Variable cardio binary

3. PROPOSED SYSTEM
Our system is an app-basedmachinelearning application
trained on a dataset from Kaggle. The admin inputs the
required attributes to get the prediction for that patient. The
model will determine the probability of heart disease. The
android application displays the result of the forecast.
We have tested the following three algorithms:
1. Logistic Regression
2. Naïve Bayes
3. Support Vector Machines
The algorithms have been trained and tested using the
dataset obtained from Kaggle. 80% of data is used for
training, while the rest is used for testing. Additional
processing is done on the dataset to make the training
process efficient and make the model accurate. The
algorithms were tested and judged based on their accuracy,
and Logistic Regression was the most accurate. Hence, we
selected it for the main application. Flask API is used for
connecting the python backend with the androidapplication.
We have used SQL to store the patient's data and Admin user
accounts.
Input: Data Set, Patient Data
Output: Prediction of Cardiovascular Disease
Figure -1: Architecture Diagram
4. PROCESS
Figure -2: Training Process
4.1 Data Preprocessing
1. Data Cleaning: The dataset obtained from Kaggle is
not perfect; first, wecheck if the dataset hasanynull
values. If there are any empty columns, we will
remove the records from the dataset. The dataset
contained no null values, so we deleted no record.
We also removed the blood pressure values, which
were unrealistic from the dataset.
2. Removing Outliers: The next step we took was to
detect and remove outliers. Outliers are the values
that look different from the other values in the data.
Below is a plot highlighting the outliers in ‘red’ and
outliers can be seen in both the extremes of data [3].
We removed height and weight values from the end
to normalize the data.Thefollowingfigure(Figure2)
shows the boxplot for the data used to detect
outliers.
Figure -3: Detecting Outliers

3. Splitting Data: After preprocessing, the remaining
data is divided into two parts, Training and Testing.
Training data is used for the Training algorithm.
Testing data is used for evaluating the algorithm.
After testing, we can calculate the accuracy of the
algorithm.
4.2 Training Model
We are using a Logistic Regression classifier for training a
model for the system. It is a statistical model used for binary
classification. We are using Logistic regression because we
have to make a binary prediction of the Target Variable,
which denotes whether a person is likely to have heart
disease or not. Once the model is trained, we move on to
evaluate it.
5. IMPLEMENTATION
Once the model is trained, weneed a way to give itthedatato
process and get the predicted result. We have used an API to
access the model from outside the system to do this. Wehave
built a REST API using the FLASK framework.
An API, or application programming interface, is a collection
of directions that define how programs or devices can link to
and communicate with each other. A REST API is an API that
fits the design principles of the REST or representational
state transfer architectural style. For this reason, REST APIs
are sometimes referred to as RESTful APIs [3].
Flask is a micro web framework written in Python. It is
categorized as a microframework because it does not need
particular tools or libraries [4].
Figure -4: API Connection
By implementing the API, we can access the model and its
methodsand functions fromothersystemsinthenetwork.As
this creates a security concern for the design, we have
implemented Authentication for the API using OAuth 2.0.
OAuth 2.0 focuses on client developer transparency while
providing detailed authorization flows for web applications,
desktopapplications,mobilephones,andlivingroomdevices.
IETF OAuth Working Group does the development of
specifications and extensions [5].
We have decided to make an android application using the
API.
6. WORKING
6
Figure -5: Implementation Process
6.1. IP Address of the Server
The admin needs to enter the server'sIPaddress,hostingthe
FLASK RESTful API. This step is required because the server
is hosted on the local network whose IP address changes
periodically.
6.2 Admin Login
Once the application successfully connects to the APIserver,
the admin must authenticate with the API. When the user
enters the valid credentials, the API server sends a Token
that is valid for 90 mins and is required for accessing other
endpoints of the API.
6.3 Getting Patient data
Once the admin is successfully logged in, we can send the
patient’s personal and medical details. The data is stored in

the database, and the medical information is then used to
indicate whether the patient has CVD or not. The prediction
is displayed on the next screen to the user and stored in the
database.
7. RESULTS
The heart disease prediction applicationevaluatesthe risk of
whether a person has heart disease or not successfully. The
system has 73% accuracy in diagnosing the heart disease.
The following diagram shows the ROC curve of the model.
An ROC curve (receiver operating characteristic curve) is a
graph showing the performance of a classification model at
all classification thresholds. This curve plots two
parameters- True positive rate and False positive rate [6].
Figure -6: ROC curve
8. CONCLUSION
We tested 3 algorithms and selected Logistic Regression. We
chose it because it had the highest accuracy of all three. The
False negatives of this model is also low compared to the
other models which can beseeninthefollowingfigure which
is a confusion matrix for the logistic regression model.
Figure -7: Confusion Matrix
For creating this application, we have used Python for
training the model and implementing the Flask API, Java for
creating the android application. By using these tools, we
have created a user-friendly UI which can be used by novice
users very easily and no technical knowledgeisneeded. This
will help users with a preliminary prediction about their
heart condition. Since Heart Diseases are increasing rapidly
in the world this application can be used for fast diagnosis
which can have a profound impact on the world.
ACKNOWLEDGEMENT
We have great satisfaction in presentingthereporton"heart
disease prediction using Machine Learning". We take this
opportunity to express my sincere thanks to my guide, Prof.
Rovina Dbritto, for providing the technical guidelines and
suggestions regarding the line of this work. We want to
convey our gratitude for her constant encouragement,
support and guidancethroughouttheproject'sdevelopment.
We are grateful to Dr J. B. Patil (Director, UCOE, Kaman) and
Prof. Yogita Mane (HOD, Information Technology, UCOE);
our project would not have shaped up withouttheirsupport.
We wish to express my deep gratitude toward all my
colleagues at Universal College of Engineering, Kaman, for
their encouragement.
REFERENCES
[1] https://www.techtarget.com/searchenterpriseai/defi
nition/machine-learning-ML
[2] https://www.who.int/news-room/fact-
sheets/detail/cardiovascular-diseases-(cvds)
[3] https://www.ibm.com/in-en/cloud/learn/rest-apis
[4] "Flask Foreword"
https://flask.palletsprojects.com/en/2.1.x/foreword/
[5] https://oauth.net/2/
[6] https://developers.google.com/machine-
learning/crash-course/classification/roc-and-auc

Heart Disease Prediction using Machine Learning

More Related Content

Similar to Heart Disease Prediction using Machine Learning (20)

More from IRJET Journal (20)

Recently uploaded (20)

Heart Disease Prediction using Machine Learning