Int. J. Signal and Imaging Systems Engineering, Vol. 0, No.

00, 201X 1

Artificial Intelligence Techniques for the Early

Diagnosis of Cardiovascular Disease in High-Risk
Patients: A Review

Atyaf Alshehri
Departmenaut of Computer Science,
King Khalid University,
Abha , Saudi Arabia
Fax: +1 E-mail: 445816229@[Link]

Abstract: Cardiovascular disease (CVD) is recognized as the leading cause of global

morbidity and mortality; hence, timely and precise detection is of utmost importance for
successful treatment and prevention of complications. where Artificial Intelligence (AI)
methods, in particular machine learning and deep learning, have demonstrated potential
for enhancing the diagnosis of CVD recently. This literature review would assess the
standards and effectiveness of AI techniques in the diagnosis of cardiovascular disease
for high risk individuals. Different AI techniques have been applied and presented with
different outcomes of success. AI methods provide faster and more accurate recognition
due to the procedural automation of data analysis and pattern recognition, while manual
interpretation of medical images is time-consuming and inaccurate. The review mainly
focuses on the strong points and the weak points of the field, as well as on the current
obstacles, giving the possibility of progress and development. Machine learning and deep
learning techniques in CVD diagnosis greatly improve health results.

Keywords: Cardiovascular disease, Artificial intelligence, Machine learning, Deep

learning, Data analysis, Risk assessment, Early detection, Healthcare.

1 Introduction high risk for CVD. Several studies have been carried
out to examine and assess whether different artificial
Cardiovascular disease (CVD), with all its face problems, intelligence methods are any good for detecting the
unquestionably remains the core of mortality in the CVD disease. These methods have been employed using
whole world and really contributes to the healthcare various datasets for the for the purpose of appraisal.
issue. A report of World Health Organization(WHO) These studies were performed and showed the fact
of 2019 based on cardiovascular diseases revealed that that AI tests can greatly improve the precision of the
CVDs were responsible for 38% of the total number of diagnosis of CVD disease.
deaths worldwide of more than 17 million adults below Traditional tools for CVD detection are based on
the age of 70(WHO page) . Unknown to the exact cause the manual interpretation of the medical images: ECGs,
of CVD, there are so many instances and/or paths that medical scans, and echocardiograms. Entail a lot of
one can possibly get this. Also, the risks of having heart subjective, unsustainable, and prone-to-error human
problems are other factors that support the existence of obstacles[3]. AI techniques are a means of avoiding the
heart disease. There are some common issues that may problem that is related to a deficit of human power
be a major factor in developing CVD disease, such as to process and analyze all medical data that exists by
smoking, high blood pressure, obesity, high cholesterol, automating this process and outlining the main features
diabetes, family history, and age. and patterns. This method can be performed faster
The early and precise diagnosis of heart disease, and more proficiently than an ordinary human observer
together with timely and accurate treatment, are truly and can perceive even more complicated relationships
important in order to treat the patient successfully and between parameters. Through implementing ML and DL
avoid the occurrence of further complications [1]. In approaches for the diagnosis of cardiovascular diseases,
addition to the apparent effectiveness of AI methods the researchers can provide high levels of accuracy and
such as machine learning and deep learning, this trend exactness in identifying heart disease[4].
sees an increasing interest in finding and diagnosing This review of the literature is undertaken to assess
cardiovascular diseases[2]. AI-based prediction tools are the standards of AI methods employed for the diagnosis
commonly attributed to high accuracy and sensitivity, of cardiovascular illness. This review is structured
which can in turn predict early warning signs and around some major sections. The next part covers
offer timely interventions, in most cases for those at ultramodern AI techniques for the determination of
2 author

cardiovascular diseases using machine learning and deep diseases in cardiology as it classifies data, determines the
learning. In the third section, we will focus on the early disease character on quantitative or semantic grounds,
diagnosis of cardiovascular disease for high-risk people, structures the decision-making process, provides the
like people with diabetes, hypertension, or those who best classification of the image for the diagnosis of
have a family history of the disease. The fourth part cardiovascular abnormalities, determines the pressure
delivers a lengthy overview of the investigations, and of the heart, and also assesses arrhythmia based on a
the conclusion gives the main deductions and points to single-channel electrocardiography. Thus, deep learning
the probable influence of AI methods in enhancing early has created a high speed of development and can have
diagnosis and treatment of cardiovascular disease. a very great impact on how cardiology is designed as
individuals, medicine is utilized, and many forms of data
are integrated. [8].
2 Modern AI Techniques in detecting
cardiovascular
3 Early Detection of cardiovascular Disease
Artificial intelligence has created a new position in High-Risk Individulas
for the health sector with advanced possibilities
to accurately diagnose and predict certain diseases. The statistics point to cardiovascular diseases
Artificial intelligence has proven itself as an integral everywhere as the key causative of fatalities across
instrument that is being used to diagnose diseases in the globe. As per the World Health Organization,
their early stages, find risks, and, thereafter, help doctors those who suffers from diabetes, high pressure or their
make decisions about starting a treatment early[5]. family history of cardiovascular disease have the higher
hazard of grappling with this disease. On the top
2.1 Role of Machine learning in cardiovascular of this, the unhealthy lifestyle plays the main role,
detection e.g. Tobacco smoking, inappropriate diet, insufficient
physical activity, and obesity are the major risk factors
Cardiovascular disease detection and analysis are now of the cardiovascular diseases(WHO page) . Therefore,
transformed with the use of a super-powerful technology prompt identification and care for cardiovascular
called machine learning. This technology enables the diseases are both of key factors in the long run, which
highly accurate detection and prediction of heart and can help to prevent the course of the illness and improve
vascular conditions. Machine learning (ML) involves clinical outcomes.
the general purpose of the ability to learn without
explicit program rules and the receipt and further 3.1 Detecting Cardiovascular Disease among
development of experience that comes with time. ML Diabetes Patients
includes supervised, unsupervised, and reinforcement
learning. Supervised learning includes labeled data that Detecting CVD among diabetic patients is so important
is used to teach the algorithm the mapping rule for because of their high risk of cardiovascular problems.
inputs to outputs, whereas unsupervised learning is done It is key to identify and check on CVD risk markers
to identify the hidden patterns in data and group them and proper control of blood sugar in the group of
without the labeled data, while the fundamental basis of patients in consideration of their early diagnosis and
reinforcement learning is the deployment of a computer consequent prevention[9].The main objective of Dinh et
program that interacts with a certain environment with al. is an assessment of the applicability of supervised
feedback[6]. machine learning algorithms for the identification of
ML is significant in cardiovascular disease prediction those people who have the highest probability of being
and diagnosis because of the many data attributes diagnosed with diabetes and having a cardiovascular
being analyzed to obtain results accurately, comparing disease using lab test and survey answer data. The
algorithms for precision, and predicting the possibility team, whose training was based on the NHANES
of raising efficiency, reliability, and accuracy in CVD data, used logistic regression, support vector machines,
detection[7]. random forest, and the gradient boosting method. As
a result of the research, XGBoost boosted performance
2.2 Role of Deep learning in cardiovascular in cases of no lab data with an AU-ROC score of
detection 83.1% and cases of lab data with an AU-ROC score
of 83.9%, respectively. It states that ML technology
The The work on deep learning starts to be recognized as can even produce an instant life-saving diagnosis for
an important dimension of the field of machine learning the demographic that is prone to diabetes and heart-
and to be used for such tasks as diagnostic imaging, related diseases.[10]. AI was employed by researchers
image analysis and diagnosis, and phenotyping on a Kang Zhang et al. and included a system for chronic
molecular level, which can be categorized into shelf- kidney disease (CKD) and type 2 diabetes mellitus
supervised, unsupervised, and reinforcement modes. (T2DM) detection and prediction from retinal fundus
Deep learning is very useful for the diagnosis of images. Deep neural networks were trained and validated
 short title 3

for their performance. A benchmark and a public dataset gateways to atherosclerosis and heart disease is high
were provided for future research. Where The AI system blood pressure, also called hypertension, which is very
showed a high degree of diagnostic accuracy in all the common and affects a vast number of people in the
cases it considered. The performance of the system was world[14]. Yaoling Wang et al. They discussed detecting
measured by fitting linear models to the longitudinal the widened aortic vessel by AI in different positions and
cohort data and calculating the rate at which T2DM predicting severe blood pressure. They have evaluated
and CKD progressed and grew over time. This study the risk level and diagnostic efficacy outcomes of aortic
thus appraises both the identification of nondiabetic diameter by using regression and receiver operating
retinopathy patients without symptoms utilizing fundus curves. A data analysis found that the middle of the
pictures and the impact on the function of patients with descending aorta is more likely to cause abnormal blood
and without diabetic retinopathy with asymptomatic pressure behavior, which has a powerful link with the
T2DM[11]. With the use of the PyCaret library, Rejath following factors: rising blood pressure, poor BP control,
Jose and his colleagues did research to show how they and masked hypertension. Then AI is able to locate the
can predict the level of diabetes and the future reading heart defect because it can detect poor blood pressure in
of hospital admissions. It highlights a group of patients the thoracic-aortic subset[15]. The study by Feras Hatib
who are susceptible to issues and makes machine learning et al. Discuss the use of machine-learning algorithms
algorithms with high precision possible. where This to predict hypotension in clinical settings using high-
study adopts the Light Gradient Boosting Machine fidelity arterial pressure waveform analysis. where
classifier and other advanced modeling techniques to Hypotension, defined as mean arterial pressure less than
improve the level of interpretation among other features 65 mmHg, is associated with subsequent myocardial
where necessary. This research examines the usefulness of infarction and acute kidney injury after surgery.
machine learning in the prediction of diabetes in regards The algorithm developed with machine learning by
to outcomes and readmission rates and establishes researchers identified 60% of patients with hypotensive
influential parameters such as accuracy, precision, recall, events 15 minutes before observed hypotension with
and area under the curve of ROC. It assesses the 88% sensitivity and 87% specifity[16].Desmond Chuang
reliability of the machine learning models and spots the Kiat Soh et al. presented in their research the
factors predictive of diabetes and readmissions[12]. By development of an automated diagnostic tool for
utilizing advanced and state-of-the-art machine learning hypertension (HPT) using electrocardiogram (ECG)
(ML) techniques, Konstantia Zarkogianni et al. looked signals and a Convolutional Neural Network (CNN)
into the prediction of particular cardiovascular diseases model. The CNN model was trained with preprocessed
(CVD) in people with type 2 diabetes mellitus (T2DM). ECG signals from established databases, and the
The research approach was targeted to mitigate the developed model afforded a high classification accuracy
problem of the imbalanced data source through the of 99.99%. The model’s robustness in figuring out
use of novel ensemble techniques such as hybrid HPT was proven by the overall performance evaluation
wavelet neural networks (HWNNs) and self-organizing performed using many methods. The observer compares
maps (SOMs). Through constructing discrimination- the method with earlier research that used deep
based responses, the test found the model’s superb learning and ADS (conventional device getting to know)
predicting ability. when compared to the binomial linear with ECG indicators,proving the effectiveness of the
regression model. The hybrid mixture of HWNN and method[17].One study is done by Jaypal Singh Rajput
SOM type models differentiated itself from the others et al. in which the problems in monitoring hypertension
by exhibiting the highest discrimination performance (HPT ) and automated detection of these patterns
of 71.48% when examining the ensemble’s AUC values. by analyzing ballistocardiography (BCG) signals are
These techniques can have clinical significance when identified, whereas the problem of hypertension is
compared to traditional statistical and machine learning one of the main health issues. Researchers are
strategies, which helps make the transference of lab interested in a technique approximately consisting
results to clinical medicine cases relatively easy[13]. of taking the processed BCG signals and applying
On the whole, these investigations add to the current a continuous wavelet transform to them that leads
knowledge on detecting and preventing cardiovascular to images and scalograms and feeding them into a
disease (CVD) among people who have diabetes with 2D CNN (convolutional neural network). The given
biomarkers, machine learning models, and indices such as model attained the highest possible classification
TyG index being assessed for their value in risk appraisal accuracy of 86.14% with a cross-validation approach
and prevention. comprised of ten folds. The application of the 2D-CNN
architecture for the first time to detect HPT using
3.2 Detecting Cardiovascular Disease among High BCG signals[18].Lastly, early diagnosis of cardiovascular
Pressure Patients disease patients who work on hypertension and vice
versa. It assists healthcare providers to detect, manage,
The discovery as well as the diagnosis of cardiovascular and prevent CVD in a timely manner.
disease in hypertensive patients is remarkable for medical
research and healthcare. Today, one of the narrowest
4 author

3.3 Detecting Cardiovascular Disease Through computerized diagnostic systems in order to make
Family History predictions concerning dyslipidemia[23]. In conclusion,
genetics is an important preventive measure for the
Genetic plays a crucial role in the development early detection of cardiovascular disease. From the
of cardiovascular disease, and its knowledge permits information, a medical practitioner can determine the
identifying CVDs in the early stages and taking genetic risk factors for CVD in an individual. With prior
preventive measures[19]. Since it is a complex disease, knowledge of the risks, they pray to take preventive
the data on its genetic aspects might be very measures and intervene early.
useful. Family history might help in determining the
risks of CVD development based on the knowledge
of potential genetic factors of the disease. Deep 4 Discussion
learning and genetic algorithms were employed in the
risk prediction of cardiovascular disease by Sharma Diabetes is closely associated with an increased risk of
K. et al. The approach identifies network weights cardiovascular complications. Therefore, some research
based on genetic algorithms, which implies employing has focused on the necessity of early detection using
multilayered feedforward neural networks for more artificial intelligence tools. In the review, we talked about
complex classification activities. Machine learning some studies that dealt with this [Link] (Dinh and
algorithms such as logistic regression, backpropagation, others) performed well by using the NHANES dataset,
support vector machines, and genetic algorithms were which was large and nationally representative. This
compared with deep neural networks in addition helped to generalize the results for a larger population
to different decision tree classification algorithms. that could be potentially affected by their findings.
With 94.34% accuracy, the identification of cardiac Although NHANES is useful data, it also has some
disease using the genetic algorithms demonstrates the limitations. It can be missing data or some bias involved
effectiveness of those techniques.[20].Zeinab Arabasadi in survey responses. So, the models may mislead the
et al. discussed in their study A hybrid method system. the study chose chose a holistic strategy for the
of diagnosing CAD using a combination of machine diagnosis of the disease by evaluating many machine
learning and data mining approaches. The proposed learning models (logistic regression, support vector
method accelerates the learning process in neural machines, random trees, and gradient boosting) to find
networks by utilizing the genetic algorithm, which the most suitable classification model for the prediction
increases performance by at least 10%. Among the four of a disease. hence the creation of the so-called weighted
ranking methods for feature selection studied, weight ensemble model by employing a synergic approach
by SVM resulted in the highest performance level. to the individual models, which is thus expected to
In contrast to the standard network, the combined enhance disease modeling effectively[10].The study by
approach of CAD prediction showed more successful Kang Zhang and others has achieved a higher and
outcomes even when tested on heart problems from other more accurate diagnosis of AI. The study in the future
datasets[21].Fergus Stafford and his colleagues shared should be evaluated and further researched in terms
the perspective that genetic testing is of great value of other clinic parameters and photos from the fundus
in finding out and managing inherited cardiac disease. through the retina for better AI system functioning[11].
In 37% of the cases resulting from genetic testing, Moreover, explorations performed by the researchers
pathogenic variants were identified, while suspicious using the PyCaret library indicate that there is an urgent
ones accounted for 8%. The research done resulted need to make further investigations and developments
in 13% of the cases getting a modified or confirmed to totally incorporate machine learning into modern
diagnosis. Conditions such as family history and age diabetes care for timely interventions and better health
onset can be used to determine these with a greater outcomes. Such a transformation focuses attention on
likelihood, including genetic background. Besides that, the role of automated machine learning in the healthcare
genomic testing should be offered to all patients with analytics realm[12].In the study presented by Konstantia
genetic disorders, which calls for the development Zarkogianni et al., which compares the performance of
of a far more diverse reference database[22].The these models with a binomial linear regression (BLR)
researchers Hamid R. Marateb et al., during their model and other commonly used machine learning
study, talk about an innovative view on dyslipidemia techniques, we found the superiority of the proposed
diagnosis using genetic and lifestyle indicators. Heart method over the binomial linear regression (BLR)
disease is associated with a genetic risk factor called model in providing reliable CVD risk scores for T2DM
dyslipidemia. The new system was superior to other patients[13].Hypertension is a well-known risk factor
classifiers in cross-validation and made a prediction for cardiovascular disease. The communication between
of adipemia. The system demonstrated capabilities in hypertension and aortic dilation, which is one of the
the model. The relationship of CETP TaqIB, ApoE, studies we talked about previously, is one of the stepping
and ABCA1 R1587K genetic factors with dyslipidemia stones in the knowledge of the effects of aorta imaging
management, lifestyle, and family background is of on CVD risk prediction and management[15]. They
scientific importance. Research was aimed at employing recommend creating other studies for the accuracy and
 short title 5

usefulness of AI-based aortic diameter measurements gap to be addressed by future [Link] expect to
in clinical practice. Likewise, the implementation of AI see more reviews discussing heart disease detection using
assessments into risk profiling models for cardiovascular DL and ML models for other populations that expose
events that link to blood pressure status is another themselves to behavioral risk factors such as tobacco
area that might need to be extensively [Link] use, unhealthy diet and obesity, physical inactivity, and
innovative approach of a deep learning technique that harmful use of alcohol.
helps in detecting high blood pressure through the
ECG signals may bring benefits to clinical settings, as
researchers indicate in a study done by Feras Hatib et 5 Conclusion
al., which shows the technique works in an effortless and
automated way to diagnose high blood pressure[16]. An This literature review has a special interest in identifying
interesting feature of the tool is that different layers of AI methods of detecting cardiovascular diseases (CVD),
the convolutional neural network are supplied with pre- and this encompasses machine learning and deep
processed signals, and the classification accuracy was learning. CVD is claiming various lives across the world.
99.99%. Moreover, further research could include clinical CVD is now the major cause of deaths all over the
trials, generalizations on various patient populations, globe. The early detection of CVD at its earliest stages
and combining ECG monitoring with other physiological is of great importance for initiating early treatment and
signals to yield better results in the management of complication prevention. AI-based prediction devices
hypertension[17].also in another study. The scientists like machine learning and deep learning have the ability
decided to use BCG signals for the detection of to assist clinical workers in increasing both accuracy
hypertension to overcome the constraints of blood and skill in the diagnosis of CVD. Such methods afford
pressure monitoring, which are almost impractical high sensitivity and accuracy in that they can aid in
without a continuous ambulatory monitoring system. the identification of CVD, thus allowing interventions
Such an approach therefore permits an uninvasive well ahead of time. Handled manually for detecting
assessment of cardiac and blood pressure functions as CVD before, it was subjective and had a lot of human
well. Regarding efficacy and accuracy, the proposed error. With the use of AI, the diagnosis will be much
2D-CNN network achieved 86.14% accuracy, which faster, if not more efficient, and results will be provided.
was greater than that of earlier machine learning Identification of CVD in those with a high risk of it, i.e.,
techniques. Also, it was realized that the model was patients with a previous diagnosis of diabetes or those
able to conduct faster computations, which improved with a family history of cardiovascular disease (CVD),
its efficiency compared to other models[18]. It can be should be a top priority. The review highlighted the
hard to underestimate the role of genetics, which can need for recognizing CVD risk factors and putting into
significantly shape the risk of a person developing CVD. place methods of regulating the course of such diseases,
Although the article about genetic algorithms written for instance, well-controlled diabetes management for
by Sharma K. et al., which uses the UCI machine diabetic patients. Suprvised machine learning algorithms
learning repository for training and testing, strengthens are being used to identify the patients with the highest
its validity, generalization of their results across various probability of being diagnosed with both diabetes
datasets is also worth to be [Link] is more, and cardiovascular disease. Moreover, they give out
right now the article by Sharma K. et al. does not early health warnings, thereby taking the necessary
take into consideration all the ethical and privacy issues measures. However, the employment of AI methods
connected with machine learning and genetic algorithms for the diagnosis of cardiovascular disease is likely to
for cardiovascular disease diagnosis and prediction[20]. prove wonderful for the realization of clinical results
This is a crucial consideration that could contribute to and the increase of healthcare processes as a whole.
the topic and the eye-opening [Link] Arabastadi Although this may be so, the issues and hindrances
and colleagues used an artificial neural network model that have to be resolved in the meantime still remain
in their experiments. They next compared the results a major concern. The review builds on the presently
of their artificial neural network models using statistical existing package of knowledge about cardiovascular
parameters, including F-score, classification variance, problems and gives scientific answers for the diagnosis
and correlation coefficients such as R2 and RMSE. of cardiovascular diseases at the early stage of their
Then, they concluded to what extent their models were pathological progression. It plays an important role in
accurate. This approach provides arguments for the the sphere of researchers, domestic health workers, and
adopted approach; thus, the research is emboldened with decision-makers in the cardiovascular field.
a rigid and reliable methodology[21].Future research
could potentially involve increasing the study population
and comprising them of equally big and different persons References
for the purpose of comparing the findings with each
other. Not only that, but more study on the function [1] Edward P. Havranek, Mahasin S. Mujahid,
of genetic testing in the management of disease and its Donald A. Barr, Irene V. Blair, Meryl S.
effect on the prevailing clinical outcomes is an important Cohen, Salvador Cruz-Flores, George Davey Smith,
6 author

Cheryl Dennison Himmelfarb, Michael S. Lauer, machine learning. BMC medical informatics and
Debra W. Lockwood, Milagros C. Rosal, and decision making, 19(1):1–15, 2019.
Clyde W. Yancy. Social determinants of risk and
outcomes for cardiovascular disease. Circulation, [11] Kang Zhang, Xiaohong Liu, Jie Xu, Jin Yuan,
132(9):873–898, 9 2015. Wenjia Cai, Ting Chen, Kai Wang, Yuanxu Gao,
Sheng Nie, Xiaodong Xu, et al. Deep-learning
[2] Mauro Chiarito, Luca Luceri, Angelo Oliva, Giulio models for the detection and incidence prediction
Stefanini, and Gianluigi Condorelli. Artificial of chronic kidney disease and type 2 diabetes
intelligence and cardiovascular risk prediction: all from retinal fundus images. Nature biomedical
that glitters is not gold. European Cardiology engineering, 5(6):533–545, 2021.
Review, 17, 2022.
[12] Rejath Jose, Faiz Syed, Anvin Thomas, and
[3] Xiaoyu Sun, Yuzhe Yin, Qiwei Yang, and Tianqi
Milan Toma. Cardiovascular health management
Huo. Artificial intelligence in cardiovascular
in diabetic patients with machine-learning-driven
diseases: diagnostic and therapeutic perspectives.
predictions and interventions. Applied Sciences,
European Journal of Medical Research, 28(1):242,
14(5):2132, 2024.
2023.

[4] Simarjeet Kaur, Jimmy Singla, Lewis Nkenyereye, [13] Konstantia Zarkogianni, Maria Athanasiou,
Sudan Jha, Deepak Prashar, Gyanendra Prasad Anastasia C Thanopoulou, and Konstantina S
Joshi, Shaker El-Sappagh, Md Saiful Islam, and Nikita. Comparison of machine learning
SM Riazul Islam. Medical diagnostic systems using approaches toward assessing the risk of developing
artificial intelligence (ai) algorithms: Principles and cardiovascular disease as a long-term diabetes
perspectives. IEEE Access, 8:228049–228069, 2020. complication. IEEE journal of biomedical and
health informatics, 22(5):1637–1647, 2017.
[5] Kanadpriya Basu, Ritwik Sinha, Aihui Ong, and
Treena Basu. Artificial intelligence: How is it [14] Crystal M Ghantous, Layla Kamareddine, Rima
changing medical sciences and its future? Indian Farhat, Fouad A Zouein, Stefania Mondello,
journal of dermatology, 65(5):365–370, 2020. Firas Kobeissy, and Asad Zeidan. Advances in
cardiovascular biomarker discovery. Biomedicines,
[6] Garry Choy, Omid Khalilzadeh, Mark Michalski, 8(12):552, 2020.
Synho Do, Anthony E Samir, Oleg S Pianykh,
J Raymond Geis, Pari V Pandharipande, James A [15] Yaoling Wang, Lijuan Bai, Jinrong Yang, Yichen
Brink, and Keith J Dreyer. Current applications Lu, Wenliang Fan, Zhuang Nie, Jie Yu, Kai Wen,
and future impact of machine learning in radiology. Ruiyun Wang, Linfeng He, et al. Artificial
Radiology, 288(2):318–328, 2018. intelligence measuring the aortic diameter assist in
identifying adverse blood pressure status including
[7] Najmu Nissa, Sanjay Jamwal, and Shahid masked hypertension. Postgraduate medicine,
Mohammad. Early detection of cardiovascular 134(1):111–121, 2022.
disease using machine learning techniques an
experimental study. Int. J. Recent Technol. Eng, [16] Feras Hatib, Zhongping Jian, Sai Buddi, Christine
9(3):635–641, 2020. Lee, Jos Settels, Karen Sibert, Joseph Rinehart, and
Maxime Cannesson. Machine-learning algorithm
[8] Chayakrit Krittanawong, Kipp W Johnson,
to predict hypotension based on high-fidelity
Robert S Rosenson, Zhen Wang, Mehmet Aydar,
arterial pressure waveform analysis. Anesthesiology,
Usman Baber, James K Min, WH Wilson Tang,
129(4):663–674, 2018.
Jonathan L Halperin, and Sanjiv M Narayan. Deep
learning for cardiovascular medicine: a practical
[17] Desmond Chuang Kiat Soh, Eddie Yin Kwee Ng,
primer. European heart journal, 40(25):2058–2073,
V Jahmunah, Shu Lih Oh, Ru San Tan, and
2019.
U Rajendra Acharya. Automated diagnostic tool for
[9] Christos Damaskos, Nikolaos Garmpis, Paraskevi hypertension using convolutional neural network.
Kollia, Georgios Mitsiopoulos, Danai Barlampa, Computers in Biology and Medicine, 126:103999,
Athanasios Drosos, Alexandros Patsouras, Nikolaos 2020.
Gravvanis, Vasileios Antoniou, Alexandros Litos,
et al. Assessing cardiovascular risk in patients with [18] Jaypal Singh Rajput, Manish Sharma, T Sudheer
diabetes: an update. Current cardiology reviews, Kumar, and U Rajendra Acharya. Automated
16(4):266–274, 2020. detection of hypertension using continuous wavelet
transform and a deep neural network with
[10] An Dinh, Stacey Miertschin, Amber Young, and ballistocardiography signals. International Journal
Somya D Mohanty. A data-driven approach to of Environmental Research and Public Health,
predicting diabetes and cardiovascular disease with 19(7):4014, 2022.
 short title 7

[19] Michal Vrablik, Dana Dlouha, Veronika

Todorovova, Denes Stefler, and Jaroslav A Hubacek.
Genetics of cardiovascular disease: How far are
we from personalized cvd risk prediction and
management? International journal of molecular
sciences, 22(8):4182, 2021.

[20] Kratika Sharma and T Satya Kiranmai. Prediction

of cardiovascular diseases using genetic algorithm
and deep learning techniques. Int. J. Emerg. Trends
Eng. Dev, 3:1–17, 2021.

[21] Zeinab Arabasadi, Roohallah Alizadehsani,

Mohamad Roshanzamir, Hossein Moosaei, and
Ali Asghar Yarifard. Computer aided decision
making for heart disease detection using hybrid
neural network-genetic algorithm. Computer
methods and programs in biomedicine, 141:19–26,
2017.
[22] Fergus Stafford, Neesha Krishnan, Ebony
Richardson, Alexandra Butters, Sophie Hespe,
Charlotte Burns, Belinda Gray, Caroline Medi,
Natalie Nowak, Julia C Isbister, et al. The role of
genetic testing in diagnosis and care of inherited
cardiac conditions in a specialised multidisciplinary
clinic. Genome Medicine, 14(1):145, 2022.
[23] Hamid R Marateb, Mohammad Reza Mohebian,
Shaghayegh Haghjooy Javanmard, Amir Ali
Tavallaei, Mohammad Hasan Tajadini,
Motahar Heidari-Beni, Miguel Angel Mañanas,
Mohammad Esmaeil Motlagh, Ramin Heshmat,
Marjan Mansourian, et al. Prediction of
dyslipidemia using gene mutations, family history
of diseases and anthropometric indicators in
children and adolescents: the caspian-iii study.
Computational and structural biotechnology journal,
16:121–130, 2018.