Vol.:(0123456789)
SN Computer Science (2020) 1:170
https://doi.org/10.1007/s42979-020-0097-6
SN Computer Science
ORIGINAL RESEARCH
Heart Disease Prediction Using CNN Algorithm
VirenViraj Shankar
1
· Varun Kumar
1
· Umesh Devagade
1
· Vinay Karanth
1
· K. Rohitaksha
1
Published online: 15 May 2020
© Springer Nature Singapore Pte Ltd 2020
Abstract
In this paper, we aim to predict accuracy, whether the individual is at risk of a heart disease. This prediction will be done
by applying machine learning algorithms on training data that we provide. Once the person enters the information that is
requested, the algorithm is applied and the result is generated. Obviously, the accuracy is expected to decrease when the
medical data itself are incomplete. We implement the prediction model over real-life hospital data. We propose to use con-
volutional neural network algorithm as a disease risk prediction algorithm using structured and perhaps even on unstructured
patient data. The accuracy obtained using the developed model ranges between 85 and 88%. We have proposed further by
applying other machine learning algorithms over the training data to predict the risk of diseases, comparing their accuracies
so that we can deduce the most accurate one. Attributes can also be modifed in an attempt to improve the accuracy further.
Keywords Machine learning · Big data analytics · Deep learning · Medical applications · Convolutional neural network
Introduction
It is reported that 50% of Americans sufer from at least
one chronic disease. Unsurprisingly, this results in 80% of
US healthcare fee being spent on chronic disease treatment.
With the raise in the living standards, the efect of these dis-
eases also increases. The USA as a whole has spent almost
$2.7 trillion per annum on respective treatments. The USA
is not the only country where large sums are spent treating
chronic diseases. In China, for example, most people die
because of chronic diseases, as reported, this accounts for
more than 85% of all deaths in the world’s most populated
country. Clearly, it is essential that early diagnosis and treat-
ment are essential, not just to save costs, but also to save
human life and improve quality of life.
Chen et al. proposed a healthcare system using smart
clothing for sustainable health monitoring. Qiu et al. had
thoroughly studied the heterogeneous systems and achieved
the best results for cost minimization on tree and simple path
cases for heterogeneous systems. Patients’ statistical infor-
mation, test results and disease history are recorded in the
EHR, enabling us to identify potential data-centric solutions
to reduce the costs of medical case studies.
With the development of big data analytics technology,
more attention has been paid to disease prediction from the
perspective of big data analysis; various researches have
been conducted by selecting the characteristics automati-
cally from a large number of data to improve the accuracy
of risk classifcation, rather than the previously selected
characteristics.
To solve these problems, the structured and unstructured
data can be combined in healthcare to assess the risk of
disease.
How Model Works?
Figure 1 depicts the various steps carried out during the
prediction of heart disease.
1. It starts with the data collection; here in this step, difer-
ent types of data mainly structured, semi-structured or
unstructured can be collected from various sources like
hospital, etc.
2. Once the data are collected, the obtained data are frst
cleaned to remove missing values and to bring under
same level of granularity, and then, the cleaned data are
classifed into training data and test dataset.
This article is part of the topical collection “Advances in
Computational Intelligence, Paradigms and Applications” guest
edited by Young Lee and S. Meenakshi Sundaram”.
* Varun Kumar
varunkumar8156@gmail.com
1
Department of Computer Science and Engineering,
JSS Academy of Technical Education, Visveswaraya
Technological University, Bangalore 56006, India