1

The Impact of Machine Learning on Healthcare Diagnostics

Abstract
The world is currently undergoing a rapid transformation in technology that will drastically
change our lives, and potentially redefine what it means to be human. Machine learning has
advanced significantly in the last few years. AI has become popular recently because of the
large-scale data processing and managing capacities of machines. The application of machine
learning in healthcare has two main domains: computer science and medical science.
Machine learning techniques have brought advancement in medical science, allowing for the
analysis of complex medical data. However, in the healthcare industry, machine learning
serves as the doctor's brain and knowledge. In this paper, we'll look at some of the benefits of
ML-based solutions and how they can be applied to healthcare. In recent years, many
conferences have been held to explore new automated technologies in medicine to provide
better health care. Several researchers are working in this domain to bring new innovations
and features. This paper will outline the main advancements of machine learning in the
automation of data analysis for patients' health records and making predictions based on the
data. It will also compare the leaps that have been made in computer-aided diagnosis, drug
discovery, and personalized medicine.

INTRODUCTION
What is machine learning?
With the ever-growing amount of data being produced every single day, traditional
programming methods are finding it increasingly difficult to structure, analyze and adapt to
the data as they offer less flexibility. Nowadays there is an increasing demand for systems
that can learn from data by finding patterns and relationships between data sets to make
better predictions. Machine learning is a branch of Artificial intelligence that enables
computers to learn from data without being explicitly programmed by someone.
It uses algorithms that can be trained and optimized to make accurate predictions when the
system encounters new data it hasn’t seen before. The process of machine learning includes
data collection, training, validation, and testing procedures in which each phase has its own
set of requirements. Machine learning combines the knowledge of data science, statistics,
 2

linear algebra, and computer science to create algorithms that can learn from data uncover
key insights from trends and solve classification problems.
The tremendous increase in the computational power of modern machines has made the
process of machine learning much simpler and helped the industry to progress fast within the
last few decades. Machine learning systems have become more ubiquitous and integrated
into every aspect of our daily life. Machine learning algorithms are behind the
recommendation systems for most shopping sites like Taobao and Amazon, news feeds
presented on your social media, tagging people from photos, and translation of speech or text
to a different language. These algorithms are becoming more powerful as more data is being
uploaded to the internet and companies are taking advantage of this to predict customer
behavior and recommend ads.

Types of machine learning

There are mainly three types of machine learning used across the industry. These are
Supervised, Unsupervised and Reinforcement learning.
Supervised learning: the most common among the three is a process of using data that is
labeled to train machine learning algorithms. Every input data is mapped to the
corresponding label before being fed into the model. The model will then learn to associate
the features to the output and selects the right parameters to make predictions for new data.
Supervised learning includes regression and classification problems. While regression
produces continuous output, classification maps the features to a discrete number of
outcomes. Supervised learning needs external supervision and guidance to provide the
desired outcome. Some examples of supervised learning include Linear and Logistic
Regression, Support Vector Machines, K Nearest Neighbor, Decision Tree, Random Forest,
and Naïve Bayes.
Unsupervised learning: is a process of clustering or grouping of data based on the similarity
between data points. As its name indicates this method of learning requires outside
intervention. The algorithm itself tries to discover patterns within the data and returns the
results. The data used to train this system is unlabeled. It means that each data will not be
mapped to a corresponding output. Some of the most common types of unsupervised learning
include K Means Clustering, DBSCAN, hierarchical Clustering, and Principal Component
 3

Analysis.
Reinforcement learning: is a process where a machine is trained to take action with regard
to a situation. The action is reinforced by rewarding the algorithm every time it makes the
decision. The agent will learn from its environment and take appropriate actions. Some of the
most important reinforcement learning algorithms include Q-learning, Sarsa, Monte Carlo,
and Deep Q Network.
Deep Learning
Deep learning is a subset of machine learning which uses perceptron or neural networks that
are made up of more than three layers. It imitates the way how our brain works to pass
signals from one neuron to the next. Each neuron in the network is just a function which
takes input from the previous layer and spits out an output. The connections between the
layers represent the weights and biases that can be optimized during training. “Deep-learning
architectures such as deep neural networks, deep belief networks, deep reinforcement
learning, recurrent neural networks, and convolutional neural networks have been applied to
fields including computer vision, speech recognition, natural language processing, machine
translation, bioinformatics, drug design, medical image analysis, material inspection, and
board game programs, where they have produced results comparable to and in some cases
surpassing human expert performance”.
Machine learning in health care
Machine learning has made its impact in the health care industry by reducing the burdens of
clinicians and doctors in managing and diagnosing patient records. It is replacing most of the
outdated and labor-intensive work of collecting, interpreting, and analyzing medical
information. It is automating the billing process and helps to increase the awareness of health
care professionals to make informed decisions about patients’ conditions. The performance
of most machine learning systems in detecting and diagnosing has already surpassed
professional experts. It has also shown promising results in providing personalized treatment
for patients, assisting in robotic surgery, organizing and classifying patients’ medical records,
drug discovery, and detection of diseases from medical images.
These algorithms have achieved high accuracy and are already deployed by most health care
providers. One thing to note here is that machine learning is a tool, and it is supposed to work
in conjunction with humans to better facilitate the process of providing high-quality medical
 4

services. Although having achieved very high accuracy, these systems cannot operate alone
as they require close supervision by humans. So, it is required to train health care
professionals by introducing them to the new algorithms so that they can better adapt to the
latest trend. Some important use cases include the remarkable achievement of Google’s deep
learning system to detect breast cancer and diabetic retinopathy from medical images which
resulted in better accuracy than radiologists. Natural language processing technologies have
also been used in interpreting unstructured patient records and charts to extrapolate important
features like patients’ medications, treatment plans, and conditions.

Figure 1 The main applications of machine learning in healthcare.

APPLICATIONS OF MACHINE LEARNING

Application of Machine Learning Models in Healthcare
The algorithms of Machine learning are “useful in identifying complicated patterns within
prosperous and huge data. This facility is especially well-suited to clinical applications,
particularly those for people who rely on advanced genomics and proteomics measurements.
It is often used utilized in numerous illnesses diagnosis and detection. In medical
applications, machine learning algorithms will manufacture higher decisions regarding
treatment plans for patients by suggestions of implementing useful health-care system”
outlines that the proliferation of patient data in hospitals and the search for new diagnostic
and therapeutic procedures have resulted in increasing demand for machine learning
applications that could offer solutions. Disease detection and diagnosis, drug development,
tailored therapy, and virtual assistants are all applications of machine learning in healthcare.
It promotes the automation of data analysis and strengthens the healthcare system.
 5

Machine learning for Personalized Medicine

One useful aspect of machine learning is delivering personally tailored treatment for patients
based on their previous medical records. The system provides medical advice to the patient
using statistical tools based on the patient’s symptoms and genetic information. The system
can be delivered in the form of an App that the patient can use to upload their symptoms and
conditions to get recommendations. This greatly reduces the time and cost of going to the
hospital to receive treatment. Supervised machine learning algorithms have been used to
create models that use computational tools to deliver treatments specifically tailored to the
individual. An example of a mobile app that provides personalized treatments is the Skin
Vision app. Patients can use this app to check whether they have skin cancer by uploading an
image of their skin looked into the use of electronic health records (EHR) in the investigation
of myocardial infractions (MIs), which usually result in a heart attack. They used EHRs and
machine learning algorithms to predict disease onset and provide personalized health scores
(risk stratification) to patients, as shown in Figure 2. They mentioned that using EHR data to
supplement traditional clinical studies is difficult with current machine learning techniques.
They proposed that relational methods such as relational functional gradient boosting
(RFGB) can learn from large datasets to help address some of these challenges. They used
the AUC-ROC curve to compare the performance of different models and concluded that the
RFGB model outperformed other tree and forest learners, Bayesian models, and SVMs.

Figure 2 Machine-Learning Systems Can Augment Current Clinical Analyses by Producing

Personalized Health Profiles Given Medical Timelines of Incoming Patients.
Machine Learning for Drug Discovery
 6

Santayana said, “Those who do not remember the past are condemned to repeat it”. This
observation applies as much to drug discovery as it does to other aspects of human endeavor.
“One constant in drug discovery is that every few years the estimated cost to develop drugs
rises further. Less than 20 years ago, developing a drug took ~12 years, cost under a billion
dollars, and the biggest challenges were failures due to efficacy or toxicity-induced attrition”.
Recently the drug industry is starting to acknowledge the impact that machine learning
models have in facilitating the drug discovery process. Machine learning will not only
improvise the manufacturing of drugs but also cut down the cost of discovering new drugs.”
Machine learning models, such as support vector machines, K nearest neighbors, naive
Bayes, random forest, and many other methods, have long been utilized for drug discovery.
However, recent interest in deep learning or deep neural networks (DNNs) for drug discovery
has catalyzed interest in machine learning in this field more broadly. DNNs have been used
in pattern recognition and machine learning, sparking their use in pharmacology and drug
discovery, and becoming a source for numerous recent reviews”. Companies like Microsoft
Project Hanover and BenevolentAI have already jumped on board to utilize machine learning
models to facilitate the drug discovery process.

Figure 3 Implementing end-to-end (E2E) machine learning models at all stages of drug
discovery.

Diagnosis of Diabetes
 7

Diabetes is a very deadly disease and a huge challenge for the public. Diabetes mellitus
begins to develop when a person consumes an extremely high level of sugar in his or her diet.
If we do not take care of this disease when it is very early on, it can develop into a very
harmful disease with many adverse effects. According to a recent study, diabetes can cause
Alzheimer's disease and blindness, and some people have experienced kidney problems due
to high blood sugar levels.

Machine Learning in Radiology

Medical imaging (radiology) is a technique used by health clinicians to diagnose diseases by
looking at what’s going on inside the body without requiring a surgery or other invasive
procedures. It has helped a lot by enabling medical experts to evaluate the situation from the
images and prescribe the necessary procedure to address the issue. It reduces the time and
resources spent on carrying out unnecessary tests and helped the doctors to make an informed
decision regarding the patient. Some of the most common types of medical images used for
diagnosing patients include CT (Computer Tomography), MRI (Magnetic Resonance
Imaging), Ultrasound, X-ray, and nuclear medicine imaging (including positron-emission
tomography (PET)).

Although medical images have been around for a long time, they require a lot of training to
be interpreted in the right manner. Even with the right amount of training, clinicians are
prone to making errors. It usually requires a consensus among different health experts to
provide the final interpretation. And this will usually result in a disagreement and lowers the
diagnosis accuracy. As a result, researchers have been working to utilize some of the best
useful aspects of machine learning to speed up the process and increase detection accuracy.
Deep learning algorithms have proven to be useful in extracting features from images which
are then used for classification, detection, and image segmentation purposes. “Recently
introduced DNN, especially CNN, has improved imaging- based classification performance
in various medical applications, including the diagnosis of tuberculosis, diabetic retinopathy,
and cancer”.
 8

Figure 4 Examples of medical image modalities and their corresponding applications.

Machine learning for Cancer Prediction

Cancer is a common but dangerous word that everyone is familiar with. Cancer is the
umbrella term for hundreds of different diseases. There are various types of cancer, but all
cancers begin when abnormal cells grow beyond a certain limit. Over the last ten to twenty
years, researchers have been working on a different type of cancer, each time providing us
with a new and accurate solution. Lung cancer, skin cancer, blood cancer, breast cancer, and
stomach cancer are the most common cancers, according to research. As a result, various
machine learning and data mining techniques are being used to make cancer predictions,
some of which are nearly entirely accurate. In the last half-decade, some optimized models
for breast cancer have been developed using a well-known prediction algorithm and
techniques. There are numerous articles and reviews available on various types of
research and forecasts regarding cancer treatment, cancer types, and cancer sub types.

Researchers conduct studies on a variety of patients from various age groups and genders
before coming to a conclusion and comprehending the findings. A sufficient number of
studies indicate that machine learning and data mining techniques can be used to predict
breast cancer patients' survival rates. In order to build predictive models for repeat
forecasting in breast tumors, an investigation takes into account actualized machine learning
techniques like "Decision Tree (C4.5), Support Vector Machine (SVM), and Artificial Neural
 9

Network (ANN)". The main goal of this work was to link affectability, specificity, and
precision to the execution of these three easily understood informational calculations. This
study assumed that the SVM model would accurately and efficiently predict the breast tumor
reoccurrence.

Computer Aided Diagnosis Systems

The computer-aided medical diagnosis system is very useful in resolving some difficult
problems because it uses a knowledge base method of problem-solving. It does assist
surgeons in correctly diagnosing a patient by utilizing a query parameter posted on the
system. The data derived from clinical practices served as the foundation of these medical
policies. The process of transferring knowledge gained through clinical practice to
programming is known as mindfulness procurement. Traditional methods, such as case-based
reasoning (CBR), will assist us in resolving health-related issues to a certain extent because
that data is stored in a database that includes clinical outcome and defensive measures. When
diagnosing a patient, a database containing information about the patient's history comes in
handy.

Figure 5 depicts how we use traditional datasets to validate the performance of the proposed
CAD system. The overall result of this method will give great confidence for further deep
diving in this direction. The Japanese Society of Radiological Technology (JSRT) provided a
set of data that could be used in performing chest radiography. More than 200 chest
radiographs were included in the database, with approximately 150 images containing
abnormal nodules and less than 100 images containing normal nodules. The original image
size was 2048*2048, but it was reduced to 60*60 for thumbnailing. The dataset is used in the
training and testing phases to compute energy and variance before being passed on to the
classification phase to compute Euclidean distance and correction metrics. Diagnosis can be
made after all of these steps.
 1

Figure 5 Computer aided diagnosis system

Figure 6 Chest radiography (Sample example of the dataset)

DISCUSSION
As illustrated in Figure 7, healthcare is one of the fastest- growing industries where machine
learning concepts and algorithms are widely used (Red: highly used and light orange: Least
used). While the healthcare sector is transforming due to the ability to record massive
amounts of information about patients, the healthcare industry is collecting an enormous
volume of data that is incomprehensible to individuals. Machine learning provides a method
for discovering examples and reasoning about information, allowing social insurance experts
 1

to provide "customized care," also known as accuracy medication. There are numerous
possible outcomes for how machine learning is used in therapeutic services, and each one is
dependent on having enough information and consent to use it. Already, in response to
external investigations, cautions and proposals for a restorative practice have been produced
and hard-coded into their product. However, because that information may come from a
variety of populations and environments, its precision may be limited. Machine learning, on
the other hand, can refine using data from that "specific environment".

Advantages of Machine Learning in Health Care

The rise in the use of machine learning in healthcare has led to a significant increase in the
amount of data and analytics available for analysis. Machine learning can help improve how
doctors prepare and deliver treatment, and lead to lower care costs and increased patient
satisfaction. Here are some of the advantages of machine learning in health care:
Improving patient care: Using modern algorithms with patient data sets and sources can
assist doctors and other medical professionals screen for illnesses with a high degree of
accuracy. AI technology can process information quicker than any person, making it a
perfect supplement to any clinician's medical profession and a highly powerful means to
acquire meaningful data. The goal is not to replace medical practitioners, but to employ AI as
clinical decision support.
Preventing and Quickly Treating Infections: Many illnesses and diseases that may take
hours or days to identify accurately using traditional approaches could be recognized in a
matter of minutes utilizing trained machines. If we can diagnose these dangerous illnesses
early, we can reduce the death and morbidity rates associated with them.
Improvised radiotherapy: To enhance and have more focused radiation treatments, a
machine learning system that can detect changes in healthy and malignant cells is being
developed. Some cancer focal sites and tumors cannot be represented using complex
equations. In medical image processing, there are various discrete variables that might occur
at any one time. Because machine learning algorithms frequently learn from a myriad of
various samples, it is often much easier to make some of the diagnoses and discover the main
parameters.
Upscaled medical data: The unprecedented growth of electronic health records has
 1

increased the amount of medical data available on patients, which may be utilized to improve
treatment. Machine learning can go through the extensive medical information offered
willingly by people can be useful to battle severe diseases and provide an opportunity to
investigate circumstances.
Diminish readmissions: Machine learning can diminish readmissions in a focused on,
viable, and entirely focused way. Clinicians can acquire intermittent bearing concerning
which patients are destined to be readmitted and how they may have the capacity to vanquish
“that risk”.
Overcome “Length-of-Stay (LOS)” in hospital: Health organizations help to lessen “LOS”
and apart from that enhance another consequence for example understanding fulfillment from
perceiving victims which are probably going to have an elevated “LOS” and guarantee that
prescribed procedures are taken care.
Decrease annual mortality: Healthcare organizations identify the true cause of death and
reduce death rates by forecasting the cause of death of a person or organism within a year of
demobilization and then coordinating victims with the appropriate caretaker, and help.

Figure 7 Machine Learning application heat map across every industry

 1

“Prognosticate propensity-to-pay”: Machine learning assists health organizations by

identifying patients who are seeking suggestions or proposals, which require financial
guidance, as well as identifying patients' circumstances and checking their background and
assisting them by providing guidelines and upper level knowledge.
“Prognosticate no-shows”: Health organization deliver exact “predictive models” to
estimate and predict with each and every appointment which are scheduled by higher
authorization, to develop patient’s health and encourage them to utilization resources
efficiently and also help them to fight the disease by providing moral support.

Challenges
Although machine learning can be highly valuable in health care, there are currently very few
examples of ML models being integrated with clinical care. One issue that may develop is
that clinical reasoning employs statistical models to detect patterns in health care data from
several institutions, but ML algorithms often take data from a small number of institutions
and are difficult to generalize. For example, when the Clostridium Difficile model was
applied to data from two medical centers, it was discovered that variables in one medical
center result in risk factors that varied from those in the other. Validating a newly suggested
ML model in a specific context takes many resources and effort. There is also the issue of
obtaining appropriate inter-disciplinary domain expertise to create an ML model from
remotely gathered stale data. Modern machine learning approaches place an emphasis on
generalization beyond a training dataset rather than generalization to distinct places. “To
overcome the issue, transfer learning mechanisms must be developed. Inspired by the idea of
transfer learning technique, one possible way is to do domain transfer which adopts a model
trained on natural images to medical image applications or from one image modality to
another. Another possible way is to apply federated learning by which training can be
performed among multiple data centers collaboratively. In addition, researchers have also
begun to collect benchmark datasets for various medical image analysis purposes”. Some of
the most common challenges for machine learning technologies to be in the health care
 1

industry are listed below:

Dataset: Sometimes a company gets stuck trying to find the correct dataset for the query
they are attempting to explain and become so picky that the project is effectively derailed.
Expertise and skill: Relatively few analytics professionals and scientists have a broad
background in machine learning technologies, and some even have a background in
healthcare. To get meaningful results for healthcare issues, we need to team up data scientists
with specialists who understand the types of questions to ask.
Getting to cause analysis: Getting to causal analysis requires experts to zero in on the
specific problem that needs to be solved, which is where healthcare expertise comes in.
Prediction is concerned with what happens next and attempting to predict or provide for that
outcome—which patients may be readmitted to the dispensary or which patients may be at
risk of a heart attack. Causality is the process of determining why something is happening in
order to make a decision that may limit or change the outcome. For example, understanding
why some Parkinson's disease patients progress more quickly so that you can target the right
biomarkers for drug development to slow or stop that progression.
Investment: Building a genuinely useful artificial intelligence and machine learning practice
requires patience and a long-term commitment. Given that reality, not every healthcare
organization will be able to justify or have the time to build an in-house machine learning
practice.
Getting to behavior change: Finally, machine learning technologies may help healthcare
companies address what has previously been a difficult problem: getting patients to improve
their behavior. A/B testing has been used by healthcare organizations as an old marketing
strategy. Try approach X with one group of patients and approach Y with another, then
choose the one that works best.

IMPLICATIONS
The future implications of Machine Learning in Health Care
Medical practitioners are increasingly concentrating their efforts on the underlying required
infrastructure to efficiently deploy ML across health care. As it currently is, IoT devices may
be integrated with ML analytics to create wearable gadgets that work in tandem with
smartphones to deliver precise real-time data that is far more valuable in monitoring a
 1

person's health. Some interesting new applications of machine learning across the health care
industry are listed below:
Predicting outbreaks: Machine Learning, in collaboration with AI, is tracking and
forecasting the spread of diseases all over the world. Our scientists now have easy access to a
wealth of data from satellites and websites. This data is used by artificial neural networks to
forecast illnesses ranging from malaria to big epidemics like the more recent COVID-19.
This stream of knowledge is especially beneficial in the case of third- world nations with
inadequate medical facilities and few to no therapeutic alternatives. A typical example can be
a machine learning tool called ProMed-Mail which monitors the state of emerging diseases
and update the status of the outbreak in real- time. It correctly predicted the outbreak of the
Ebola epidemic in 2014.
Radiogenomics: Radiogenomics can be used to map images to gene expression patterns that
could result in a better diagnosis of tumors. “Radiomics is an emerging field that converts
imaging data into a high dimensional mineable feature space using a large number of
automatically extracted data- characterization algorithms”. NLP and image recognition work
hand in hand to analyze the genetic data and make accurate predictions.

CONCLUSION
This paper has presented how machine learning techniques have been utilized in the health
care industry to handle a large amount of patient data and mitigate the cost and resources
used to analyze it. Recently there is a huge demand for these types of tools to improve the
experience of patients in hospitals and improve the accuracies of diagnoses which has proven
to be better than professionally trained doctors. It also discussed that these tools are meant to
work hand in hand with doctors and require training to implement them for the required task.
There has also been an interest in using deep learning algorithms like Convolutional Neural
Networks (CNNs) to accurately detect diseased organs from medical images which are
showing a huge promise in this field. Machine learning can be used in almost all aspects of
patient care from admission to diagnosis and prescription. It has completely revolutionized
the way doctors treat their patients by personalizing the patient’s experience and presenting
accurate predictions which are used to informed decisions. It is also showing a lot of progress
in monitoring future pandemic events to reduce catastrophes on a global scale and helps to
 1

inform citizens on how to prepare for these kinds of events. Although it’s showing a lot of
progress, there is still a lot of bottlenecks to be addressed so that these machine learning tools
can adapt to more than the environment. We must provide curated data without disruption
which will simplify the training and validation phase.

REFERENCES
1. S. Brown, "Machine learning, explained.," MIT Sloan School Of Management, 21 04
2021. [Online]. Available: https://mitsloan.mit.edu/ideas-made-to-matter/machine-
learning-explained.
2. J. Hu, H. Niu, J. Carrasco, B. Lennox and F. Arvin, "Voronoi-Based Multi-Robot
Autonomous Exploration in Unknown Environments via Deep Reinforcement Learning,"
IEEE Transactions on Vehicular Technology, vol. 69, no. 12, p. 14413–14423, 2020.
3. K. G. Dinesh, K. Arumugaraj, K. D. Santhosh and V. Mareeswari, "Prediction of
cardiovascular disease using machine learning algorithms," In 2018 International
Conference on Current Trends towards Converging Technologies (ICCTCT), pp. 1-7,
2018.
4. J. Sukanya, "Applications of Big Data Analytics and Machine," International Journal of
Engineering and Computer Science, vol. 6, pp. 21963-21967, 2017.
5. A. Maxwell, "The Use of Machine Learning in Healthcare," Influencive, 08 08 2021.
[Online]. Available: https://www.influencive.com/the-use-of-machine-learning-in-
healthcare/.
6. J. C. Weiss, S. Natarajan, P. L. Peissig, C. A. McCarty and
7. D. Page, "Machine learning for personalized medicine: predicting primary myocardial
infarction from electronic health records," Ai Magazine, vol. 33, no. 4, pp. 33-33, 2012.
8. L. D. Butler, P. Guzzie-Peck, J. Hartke, M. S. Bogdanffy,
9. Y. Will, D. Diaz and J. J. DeGeorge, "Current nonclinical testing paradigms in support of
safe clinical trials: an IQ consortium DruSafe perspective," Regulatory Toxicology and
Pharmacology, vol. 87, pp. 1-15, 2017.
10. I. Kola and J. Landis, "Can the pharmaceutical industry reduce attrition rates?," Nature
reviews Drug discovery, vol. 3, no. 8, pp. 711-716, 2004.
11. S. Ekins, A. C. Puhl, K. Zorn, T. R. Lane, D. P. Russo, J.
 1

12. J. Klein and A. M. Clark, "Exploiting machine learning for end- to-end drug discovery
and development," Nature materials, vol. 18, no. 5, pp. 435-441, 2019.
13. R. Chen, X. Liu, S. Jin, J. Lin and J. Liu, " Machine learning for drug-target," Molecules,
vol. 23, p. 2208, 2018.
14. J. Lin and K. C. Wong, "Off-target predictions in CRISPR- Cas9 gene editing,"
Bioinformatics, vol. 34, p. i656–i663 , 2018.
15. Y. Chang, "Cancer drug response profle scan (CDRscan): a deep," Sci. Rep, vol. 8, p.
8857, 2018.
16. P. Sadja, "Machine learning for detection and diagnosis of disease," Annu. Rev. Biomed.
Eng., vol. 8, pp. 537-565, 2006.
17. M. Shanthi, P. Pekka and N. Bo, "Global Atlas on Cardiovascular Disease Prevention
and Control," World Health Organization, p. 3–18, 2011.
18. World Health Organization, "Cardiovascular diseases (CVDs)," World Health
Organization, 11 06 2021. [Online]. Available: https://www.who.int/news-room/fact-
sheets/detail/cardiovascular-diseases-(cvds).
19. Anand, Ayush, Shakti and Divya, "Prediction of diabetes based on personal lifestyle
indicators," 1st International Conference on Next Generation Computing Technologies
(NGCT), pp. 673-676, 2015.
20. A. Mujumdar and V. Vaidehi, "Diabetes prediction using machine learning algorithms,"
Procedia Computer Science, vol. 165, pp. 292-299, 2019.
21. M. Kim, J. Yun, Y. Cho, K. Shin, R. Jang, H. Bae and N. Kim, "Deep Learning in
Medical Imaging," Neurospine, vol. 16, no. 4, p. 657, 2019.
22. X. Liu, K. Gao, B. Liu, C. Pan, K. Laing, L. Yan and Y. Yu, "Advances in Deep
Learning-Based Medical Image Analysis," Health Data Science, 2021.
23. Y. Xiao, J. Wu, Z. Lin and X. Zhao, "A deep learning- based multi-model ensemble
method for cancer prediction," Computer methods and programs in biomedicine, vol.
153, pp. 1-9, 2018.
24. S. Amin, K. Agarwal and R. Beg, "Genetic, neural network-based data mining in the
prediction of heart disease using risk factors," IEEE Conference on Information &
Communication Technologies (ICT), pp. 11-12, 2013.
25. I. Yoo, P. Alafaireet, M. Marinov, K. Pena-Hernandez, R. Gopidi, J. F. Chang and L.
 1

Hua, "Data mining in healthcare and biomedicine: A survey of the literature," Journal of
medical systems, vol. 36, no. 4, p. 2431– 2448, 2012.
26. M. Sendak, M. Gao, M. Nichols, A. Lin and S. Balu, "Machine Learning in Health Care:
A Critical Appraisal of Challenges," eGEMs (Generating Evidence & Methods to
improve patient outcomes), vol. 7, no. 1, pp. 1-4, 2019.
27. D. Shukla, A. Pathan, R. Shaik and V. Memon, "Exploring the possibilites of using
machine learning in health care," International Journal of Creative Reasearch Thoughts
(IJCRT), vol. 9, no. 7, 2021.
28. H. Aerts, E. Velazquez and R. Leijenaar, "Decoding tumour phenotype by noninvasive
imaging using a quantitative radiomics approach," Nat Commun, vol. 5, p. 4006, 2014.