Page 49 - AIH-1-1

Artificial Intelligence in Health AI model for cardiovascular disease prediction

a growing challenge to the public healthcare system . as diagnosis and drug prediction, wearable devices or
[4]
According to the World Health Organization, an estimated implanted sensors for the patient condition monitoring, or
17 million people die from CVDs each year, particularly an artificial intelligence (AI) system using neural network
[5]
heart attacks and strokes . The statistics are projected to (NN), genetic algorithm (GA), fuzzy logic, and many other
[25]
reach 22.2 million people by 2030, if CVDs remain the techniques for the prediction and medical diagnosis .
leading cause of death and disability worldwide . CVDs are The Tabu search algorithm is an optimization technique
[6]
no longer considered an affluent nation’s disease because that uses adaptive memory to improve the local search for
the statistics of more than 80% CVD-related deaths were global optimum performance . This search algorithm
[26]
contributed by low-and-middle-income countries [7-10] . The works using a Tabu list to prevent cycling and aspiration
current clinical guideline for primary prevention of this criteria for a globally optimal solution and to prevent the
vascular disease is to identify the asymptomatic patient repetition of solutions . The present study aims to analyze
[27]
by prognostication and early diagnosis, which help with the performance of various machine learning techniques,
the reduction of high-risk CVD patients [11-13] . The heart such as artificial NNs (ANNs), artificial NN-genetic
diseases are rapidly increasing in terms of number of cases algorithms (ANN-GA), support vector machines (SVMs),
among the young generation but are disproportionately K-means, K-nearest neighbor (KNN), and decision trees
affecting men with a mortality rate twice as high than in (DT), in yielding the best model for CVD prediction. The
women . Approximately 88,000 deaths can be traced to motivation of this research is to discover the most effective
[14]
coronary heart disease, while stroke caused about 50,000 machine learning method for detecting CVDs with greater
deaths and blood circulatory disease 49,000 deaths [15-17] . precision, sensitivity, and accuracy. The main contributions
The major risk factors associated with these vascular of this work to the literature are:
diseases are age, sex, smoking, family history, cholesterol, (i) Development of a hybrid model of ANN-GA for the
poor diet, high blood pressure, diabetes, obesity, physical early prediction and diagnosis of CVD.
inactivity, and alcohol intake . Some symptoms such (ii) Comparative analysis of CVD prediction and
[18]
as chest pain and tiredness are symptomatic indications diagnosis model using different types of supervised
aiding in the detection of CVD. However, diagnosing machine learning techniques, such as ANN, K-means,
heart diseases is a difficult task in the field of medicine KNN, and DT.
because a thorough analysis of patient clinical data and (iii) Development of an expert system-based application
health history is required. Thus, an intelligent automated program interface (API) for CVD prediction and
system for medical diagnosis can help enhance health- diagnosis.
care observation and analysis, thereby facilitating the (iv) Evaluation of the performance of different machine
administration of suitable treatment for lifesaving learning algorithms used for the prediction and
purposes [19,20] . For the prediction of heart disease, several diagnosis of CVDs.
algorithms and diagnostic models have been created to The remainder of the paper is structured as follows:
predict heart disease by considering its risk factors, such the review of related work is presented in Section 2; the
as arterial high blood pressure, high cholesterol, diabetes research methodology is described in Section 3, and the
mellitus, smoking, obesity, unhealthy diet, and a family result and discussion are presented in Section 4. Finally, the
history of heart disease [21-23] . conclusion of the research work is presented in Section 5.
The cardiac dataset obtained from the clinic is
voluminous and contains many irrelevant and redundant 2. Related works
attributes. These attributes of CVD risk factors may This section presents the strengths and weaknesses of
include inactivity, unhealthy eating, smoking, diabetes, various works conducted by other researchers, through
age, and family history. Therefore, there is a need for which we can identify the research gaps. There are different
attribute selection to reduce data redundancy and accuracy types of vascular heart diseases, including coronary
improvement using machine learning techniques to heart disease, congestive heart failure, cardiomyopathy,
predict the output from existing data . This emergence congenital heart disease, arrhythmias, deep vein thrombosis
[24]
of machine learning algorithms or intelligent automated and pulmonary embolism, and many others. Heart disease
systems is widely applicable to several fields of studies has been identified as one of the leading causes of death
for the simplification and counteraction of their natural globally. One of the reasons for heart disease-related deaths
and artificial challenges as it usually yields optimum is that the risk is not recognized at an earlier stage. Manually
performance. Such emergent technology systems in health calculating the chances of developing heart disease based
care could be an expert system for different purposes, such on risk factors is difficult, but machine learning methods

Volume 1 Issue 1 (2024) 43 https://doi.org/10.36922/aih.1746