Machine Learning Classification in Cardiology: A Systematic

Mapping Study

Khadija Anejjar

, Fatima Azzahra Amazal

and Ali Idri

LabSIV, Department of Computer Science, Faculty of Science, Ibn Zohr University, BP 8106, 80000 Agadir, Morocco

Faculty of Medical Sciences, Mohammed VI Polytechnic University, Morocco

Keywords: Heart Disease, Machine Learning, Classification Techniques, Predictive Models.

Abstract: Heart disease, a widespread and potentially life-threatening condition affecting millions globally, demands

early detection and precise prediction for effective prevention and timely intervention. Recently, there has

been a growing interest in leveraging machine learning classification techniques to enhance accuracy and

efficiency in the diagnosis, prognosis, screening, treatment, monitoring, and management of heart disease.

This paper aims to contribute through a comprehensive systematic mapping study to the current body of

knowledge, covering 715 selected studies spanning from 1997 to December 2023. The studies were

meticulously classified based on eight criteria: year of publication, type of contribution, empirical study

design, type of medical data used, machine learning techniques employed, medical task focused on, heart

pathology assessed, and classification type.

1 INTRODUCTION

Heart disease is a major global health concern and

ranks as one of the primary causes of mortality

worldwide. Although conventional approaches to

diagnosing and treating heart disease have seen

notable progress, there is increasing acknowledgment

of the potential advantages of machine learning (ML)

in enhancing medical outcomes and optimizing

cardiology practices (Hassan et al., 2022). This is

fueled by the increasing availability of diverse

medical data from electronic health records, medical

imaging, and wearable devices (Almazroi et al.,

2023).

Encompassing a range of conditions like coronary

artery disease, arrhythmia, and myocardial infarction,

heart diseases vary in complexity and require

personalized approaches considering each patient's

medical history, genetics, and environment (Collet et

al., 2021).

ML classification techniques empower heart

disease practitioners not only in disease prediction

and detection but also in patient management,

https://orcid.org/0009-0009-7299-2380

https://orcid.org/0000-0002-9008-656X

https://orcid.org/0000-0002-4586-4158

treatment, and ongoing monitoring (Esfandiari et al.,

2014).

Classification as a subset of ML (Dangare & al.,

2012; Noh & al., 2006; Seetharam & al., 2022) holds

promise for accurately predicting heart disease and

aiding doctors in making informed decisions

(Dwivedi, 2018). There are two primary types of

classification: binary classification, which

categorizes elements of a set into one of two classes,

and multi-classification, which involves assigning

elements to more than two classes (Sun, 2008). These

classification techniques are employed to identify

patterns and relationships within the data, facilitating

the categorization of patients into different risk

categories in some cases (Araki & al., 2016; Chicco

& al., 2020; Aziz & al., 2021), detecting the presence

of a heart abnormality in others cases (Chicco & al.,

2021; Hassan & al., 2022; Masetic & al., 2016;

Rahman & al., 2015), or even identifying specific

heart conditions (Juhola & al., 2018; Smole & al.,

2021).

This research utilizes a Systematic Mapping

Study (SMS) to examine ML classification in

cardiology. As defined by Kitchenham et al. (2010),

Anejjar, K., Amazal, F. and Idri, A.

Machine Learning Classiﬁcation in Cardiology: A Systematic Mapping Study.

DOI: 10.5220/0012785600003756

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 13th International Conference on Data Science, Technology and Applications (DATA 2024), pages 409-416

ISBN: 978-989-758-707-8; ISSN: 2184-285X

409

SMS establishes a framework for categorizing

research within a specific field. Notably, no prior

comprehensive mapping study has been conducted to

explore the development and current state of ML

classification in cardiology, to the authors'

knowledge. This SMS aims to: 1) Identify recent

research (1997-December 2023) on ML classification

in cardiology. 2) Evaluate and categorize the selected

literature based on eight factors: publication year,

contribution type, empirical study type, medical data,

ML techniques, medical task, heart pathology, and

classification type.

The methodology to conduct this SMS is

presented in Section 2, followed by results (Section

3), discussion of implications (Section 4), and

conclusions with future directions (Section 5).

2 RESEARCH METHODOLOGY

The systematic mapping method proposed by

Kitchenham and Charters (Kitchenham & Charters,

2007) is applied in this investigation. A mapping

study, according to Kitchenham, tries to categorize

research works in accordance with a set of

predetermined criteria and discover the research

trends associated with a certain topic (Kitchenham &

al., 2010). The following five steps make up the

utilized mapping process: defining the mapping

questions, selecting studies, extracting data,

summarizing data, and conducting a thorough search

for candidate articles.

2.1 Mapping Questions

This mapping investigation resulted in the

formulation of eight mapping questions (MQs). The

MQs and their major motivating factors are listed in

Table 1.

Table 1: Mapping questions.

ID Mapping question Motivation

MQ1

What are the years and

venues of publication

of the selected studies?

Track publication

trends and venue

MQ2

What types of

contributions were

presented in the

selected studies?

Analyze study impact

on knowledge and

practice advancement

MQ3

What research

approaches did the

selected studies adopt?

Categorize the

research approaches

that were used in the

selected studies

MQ4

Which medical tasks

received the most

attention in the selected

studies?

Identify the most

studied cardiology

tasks

MQ5

Which heart disease did

the studies focus on?

Identify prevalent vs.

less explored heart

diseases

MQ6

What types of data

were used to conduct

experiments in the

selected studies?

Analyze data type

usage (requirements

and limitations)

MQ7

What type of

classification was used

in the studies?

Identify the employed

classification types

MQ8

What ML techniques

were used in the

selected studies?

Identify dominant

ML techniques

2.2 Search Strategy

To identify relevant publications addressing the

research questions in Table 1 on ML classification in

cardiology, seven electronic databases were searched:

IEEE Xplore, DBLP, ScienceDirect, ACM Digital

Library, PubMed, Springer Link, and Google

Scholar. These choices align with prior systematic

reviews in this domain (Amazal & Idri, 2019; Idri &

al., 2018; Idri & al., 2015; Kadi & al., 2017; Kadi &

al., 2019).

The search focused on articles published between

1997 and December 2023, utilizing a comprehensive

search string targeting titles, abstracts, and keywords.

This strategy ensured inclusion of the most recent

version of each study and avoided duplicates.

The entire search string set was created in the way

that is described below.

((cardi* or heart* or vascular or arter* or coronary

or myocardial) and (defect* or disease or failure or

abnormal) and ("machine learning" or ML) and

(classif*) and (model or method or technique or

algorithm or rule or tool or framework or approach)).

The search strategy targeted relevant articles

using titles, abstracts, and keywords in the

aforementioned libraries. Only the most recent paper

for each study was included, avoiding duplicates from

various publication channels.

2.3 Study Selection

Inclusion and exclusion criteria were applied to

identify relevant articles addressing the research

questions in Table 1.

DATA 2024 - 13th International Conference on Data Science, Technology and Applications

410

2.3.1 Inclusion Criteria

 Studies aiming to predict heart diseases using

ML-based classification or to enhance that

process

 Studies aiming to compare different techniques

for predicting heart diseases using ML

classifiers

 Papers on the detection of other diseases

directly related to heart diseases (symptoms of

a heart disease or causes)

2.3.2 Exclusion Criteria

 Papers that center on predicting a variety of

diseases, alongside heart diseases

 Papers predicting potential heart disease

symptoms without explicitly targeting heart

disease detection

 Papers employing classification techniques

exclusively for the purpose of feature selection

 Papers utilizing classification techniques

unrelated to ML

 Duplicate publications (only the most complete

version is included)

 Other Systematic Mapping Studies (SMS) or

systematic literature reviews

Applying the previously described criteria, the

candidate papers were assessed, which included an

examination of their abstract, title, and in some cases,

the entire content. Subsequently, they were

categorized as either "included" or "excluded".

2.4 Data Extraction Strategy and

Synthesis Method

The extraction of data from all selected publications

addressed the research questions in Table 1. A

standardized form (Table 2) guided this process. The

extracted data were then analyzed for each question

using a narrative synthesis approach, supplemented

by relevant visuals (tables, graphs, etc.).

Table 2: Data extraction form.

Data Extractor

Paper Identifier

Author(s) Name(s)

Paper Title

(MQ1) Publication Year and Channel

(MQ2) Contribution Type (Tool, Algorithm, Model,

Framework, Metric, Comparison, Validation, Other)

(MQ3) Research Approach (Solution Proposal ,

History-based Evaluation, Case Study, Theory,

Experiment, Other)

(MQ4) Medical Task Assessed (Screening, Diagnosis,

Treatment, Prognosis , Management, Monitoring)

(MQ5) Heart Disease Studied (Arrhythmia, Coronary

Artery Disease, Cardiac Arrest, Myocardial

Infarction, Dilated Cardiomyopathy, Valvular Heart

Disease, Other)

(MQ6) Type of Data Used (Patient Medical

Characteristics, Medical Images, Electronic Health

Records, Physiological Signals, Wearable Devices

Data, Other)

(MQ7) Type of Classification Used (Binary

Classification, Multi-class Classification)

(MQ8) Machine Learning Techniques Utilized

3 RESULTS AND DISCUSSION

The findings of our mapping study in relation to the

Table 1 questions are discussed in this section.

3.1 Overview of the Selected Studies

Figure 1: Overview of the selection process and its results.

The search across seven databases yielded 15,345

candidate publications (Figure 1). After applying

inclusion/exclusion criteria and reviewing full texts,

titles, abstracts, and keywords, 715 relevant studies

were selected. Reference lists of included studies did

not yield additional relevant publications.

3.2 Publication Trends and Venues

(MQ1)

We examined how often ML classification appeared

in cardiology studies over time (Figure 2).

Publications rose steadily from 1997 to 2023. The

jump after 2016 (93% of studies) suggests growing

interest and use of ML in cardiology research.

IEEE

xplore

N=438

DBLP

N=42

Science

Direct

N=336

ACM

Digital

Library

N=361

PubMed

N=805

Springer

Link

N=5053

Google

Scholar

N=8310

Selected Papers: 715

Relevant Articles: 715

Relevance Assessment

Inclusion Criteria (CI) Exclusion Criteria (CE)

Candidate Documents: 15,345

Machine Learning Classiﬁcation in Cardiology: A Systematic Mapping Study

411

Figure 2: Publication trends of the selected studies.

Over half (57%) of the 715 studies were journal

articles, while conferences presented 32%. The rest

were chapters, symposia/workshops (each under

10%), and a single report (Figure 3). We found most

journals on Google Scholar, PubMed, or

SpringerLink, while conference papers were on IEEE

Xplore or ACM Digital Library. Chapters were

mainly on SpringerLink, and symposia/workshops

and reports were found on ACM Digital Library and

Google Scholar, respectively.

Figure 3: Publication sources and venues.

3.3 Contribution Types (MQ2) and

Research Approaches (MQ3)

As depicted in Figure 4, the selected papers employed

five primary research approaches: history-based

evaluation, solution proposals, case studies,

experiments, and surveys. Out of 715 studies, 693

were based on history-based evaluation, and 623 were

solution proposals. In contrast, there were only 12

case studies, 5 experiments, and one survey.

Figure 4: Research approaches used in the selected studies

and their contribution types.

Most studies (around 61-67%) focused on

developing new techniques, regardless of the research

approach used (history-based evaluation, solution

proposal, case study). Validation was the primary

focus for experiments (80%) while case studies were

more balanced between developing and comparing

techniques (all around 40%). There were some

overlaps, with studies often combining approaches

(e.g., history-based evaluation and solution proposal)

and contribution types (e.g., comparing and

validating new techniques).

3.4 Medical Tasks (MQ4) and Heart

Diseases Studied (MQ5)

Figure 5: Medical tasks and heart pathologies studied.

Figure 5 shows that most research focused on

screening (34%) and diagnosis (31%) of heart

diseases, followed by prognosis (24%). Other tasks

like monitoring, management or treatment were less

common (5% total). Interestingly, some studies (6%)

tackled multiple tasks simultaneously.

For heart conditions, a third (36%) didn't specify

a particular type. Coronary artery disease (18%) and

arrhythmias (23%) were the most studied, followed

by myocardial infarction (9%) and dilated

cardiomyopathy (8%). Less common conditions (3%

total) included cardiac arrest (CA), valvular heart

disease (VHD), and others. Some studies (3%)

explored multiple conditions at once.

100

150

200

1994 2004 2014 2024

publications

year

20%

40%

60%

80%

100%

Publication Types and Venues

Journal papers Conference papers

Chapters symposium

Workshop report

DATA 2024 - 13th International Conference on Data Science, Technology and Applications

412

Figure 6: Medical task and heart pathology correlation.

As shown in Figure 6, studies concentrating on

detecting arrhythmias primarily focused on the

diagnosis task, aiming to identify specific types

among various based on symptoms. For CAD, MI,

and DC, the notable tasks associated with them were

prognosis, diagnosis, and screening, each in close

percentages.

Studies where heart pathologies were unspecified

predominantly emphasized screening, reflecting

uncertainty among professionals about the specific

heart pathology they seek.

3.5 Types of Data Employed (MQ6)

The selected studies utilized various data types:

physiological signals (29%), electronic health records

(10%), patient characteristics (7.8%), medical images

(6.4%), wearable data (0.6%), and others (1.8% -

gene expression, voice recordings, etc.). While Table

3 focuses on single data types, table 4 focuses on

studies combining different data types.

Table 3: Data types used in the selected studies.

Type of data Description # of

studies

Physiological

Signals (PhS)

Data from monitoring

devices (heart rate,

blood pressure, ECGs,

etc.)

212

Electronic Health

Records (EHR)

Digital patient records

containing diagnoses,

medications, lab

results, etc.

Patient Medical

Characteristics

(PMC)

Demographic, medical

history, lifestyle, and

relevant health details

Medical Images

(MI)

visual data from X-

rays, MRIs,

ultrasounds, CT scans

Wearable Devices

data (WD)

Data from wearable

technology monitoring

activity metrics

Other types (OT) Genetic data, omics

data, or unspecified

data type

Not reported

Refers to situations

where the type of data

used is not reported

Table 4 shows a trend towards combining data in

arrhythmia research. A significant number of studies

(175) combined electronic health records with patient

characteristics. Medical images were frequently used

with most data types. Wearable devices showed

promise, with 17 studies combining their data with

physiological signals and medical characteristics.

Some studies even ventured into using three or four

data types together. Importantly, physiological

signals were the most used data, especially when

studied alone, highlighting their significance in

arrhythmia diagnosis.

Table 4: Data types combinations.

Data types combinations # of studies

EHR+PMC 175

PMC+MI 14

PhS+MI 13

PhS+PMC 13

PhS+WD 13

EHR+MI 7

EHR+PhS 4

PMC+WD 4

EHR+PhS+PMC 23

EHR+MI+PMC 10

EHR+PhS+PMC+MI 2

3.6 Classification Approach (MQ7)

Figure 7 shows trends in heart disease classification.

Binary classification, identifying presence or absence

of any heart disease, dominates (65.73%, 470

studies). Multi-classification for specific disease type

identification follows (28.67%, 205 studies). A small

portion (2.94%, 21 studies) uses a hybrid approach

for both presence and specific type. Notably, 2.65%

(19 studies) lack classification details.

Figure 7: Classification approaches employed in the

studies.

100

150

Diagnosis Screening

Prognosis Monitoring

Managment Treatment

Binary

66%

Multi-class

29%

Binary and multi-class

Not reported

Machine Learning Classiﬁcation in Cardiology: A Systematic Mapping Study

413

3.7 ML Techniques Used to Handle

Heart Diseases (MQ8)

ML techniques usage in the selected studies revealed

a predominance of multiple technique applications

(75.62%) for comparison and model building, with

some studies exploring up to 13 distinct techniques

(Garg et al., 2022; Guo et al., 2023; Anton et al.,

2021; Swathy et al., 2022). A smaller portion

(23.92%) focused on a single technique, and a small

minority (0.28%) developed entirely new algorithms.

Figure 8: ML techniques used in the selected studies.

Figure 8 shows the use of several ML techniques

in the selected studies. As can be seen, traditional ML

models (74.4%) were the most common, followed by

ensemble techniques (63.1%). Support Vector

Machine (SVM) featured prominently (46.2%), while

Neural Networks/Deep learning came fourth

(35.2%). Other techniques were also used (e.g., tree-

based methods: 28.7%). This variety highlights the

widespread use of diverse ML approaches.

Similarly, Table 5 details the traditional ML

techniques that were frequently used. KNN was the

most applied (26%), followed by Logistic Regression

(24.4%) and Naive Bayes (20.8%). K-Means and

Genetic Algorithms were less frequent (2% and 3%).

Table 5: Traditional machine learning techniques.

Traditional ML models # of papers Total

KNN (k-Nearest Neighbours) 186

532

LR (Logistic Regression) 174

NB (Naive Bayes) 149

GA (Genetic Algorithm) 14

K-means 13

Table 6 showcases Random Forests as the most

dominant ensemble technique (39.58%). XGBoost

(10.07 %,) and AdaBoost (7.28%) followed at a

distance. The inclusion of other techniques like

CatBoost, Bagging, Voting, and Stacking highlights

the variety of ensemble methods used in the research.

Table 6: Ensemble techniques.

Ensemble technique # of papers Total

RF 284

451

AdaBoost 52

XGBoost 72

Bagging/ Voting/ Stacking 32

CatBoost 11

Table 7 highlights SVM dominance (46%) in ML

approaches. The classic SVM reigns supreme (95%),

with minimal use of other variants (RBF-SVM:

1.54%, Quadratic SVM: 0.42%, and a few studies

employing even less common variations).

Table 7: Support vector machines (SVM) and its variants.

Support vector machines (SVM)

and its variants

# of papers Total

SVM 312

330

RBF-SVM 11

QSVM (Quadratic SVM) 3

Incremental SVM/ SVM

PEGASOS

LSTSVM (Least Squares Twin

SVM)/ KSVM (Kernel SVM)

Table 8 showcases Artificial Neural Networks

(ANNs) as the leading neural network technique

(17.32%). Multi-Layer Perceptrons (MLPs) and

Convolutional Neural Networks (CNNs) followed

closely (7.52% and 6.84% respectively). While

techniques like Deep Neural Networks (DNNs),

Recurrent Neural Networks (RNNs), were less

frequently present (combined total under 3.35%).

Table 8: Neural network and deep learning models.

Neural networks and deep

learning models

# of papers Total

ANN (Artificial Neural

Networks)

124

252

MLP (Multi-Layer Perceptron) 54

CNN (Convolutional Neural

Network)

DNN (Deep Neural Network) 14

RNN (Recurrent Neural Network) 6

Echo State Networks/ Bi branch

network

Layer-wise Quantized CNN/

EFCN (Efficient Fully

Convolutional Network)

Bi-LSTM (Bidirectional Long

Short-Term Memory)

Table 9 highlights Decision Trees (DT) as the

dominant tree-based technique (22.38%). Decision

tree variants like CART, C4.5, C5, J48, and J4.8 were

employed in 5.73% of the studies. Notably, Extreme

532

451

330

252

205

100

200

300

400

500

600

traditional

Ensemble

techniques

SVM and

its variants

Neural

networks

and deep

learning

Tree based

other ML

techniques

number of papers

DATA 2024 - 13th International Conference on Data Science, Technology and Applications

414

Random Trees and Feature Ranking techniques

(under 0.56%) were less frequent.

Table 9: Tree-based models.

Tree-based models # of papers Total

DT (decision tree) 160

205

CART 14

C4.5 and C5 14

J48 and J4.8 13

ERT (Extreme Random Trees) 2

FR (Feature Ranking) 2

4 IMPLICATIONS FOR

RESEARCH AND PRACTICE

This study examines the use of ML classification for

cardiology, offering recommendations for

researchers, cardiologists, and care units.

A key recommendation is for researchers to

collaborate with practitioners on real-world case

studies to bridge the gap between research and

practical application.

The analysis also highlights a focus on ML for

heart disease screening, diagnosis, and prognosis.

Further research is needed for treatment, monitoring,

and management tasks.

The study points to a dominance of physiological

signals and electronic health records (EHR) data in

current models. More exploration is encouraged for

medical images, wearable device data, and standalone

patient characteristics.

Finally, a gap is identified in the specific heart

diseases studied. While arrhythmias and coronary

artery disease (CAD) receive attention, many other

conditions require further investigation.

5 CONCLUSION AND FUTURE

WORK

This SMS explored the use of ML classification

techniques in cardiology. Our analysis of 715 studies

revealed that:

(MQ1): A surge in research interest, particularly

after 2016, with journals as the main publication

channel.

(MQ2 and MQ3): The publications primarily

adopted solution proposal and history-based

evaluation approaches. The main contributions were

the development of new techniques, comparisons of

existing ones, and their validation.

(MQ4) and (MQ5): The selected papers mainly

focused on screening, diagnosis, and prognosis tasks.

The heart disease handled is often not mentioned, and

in some cases, the focus is on arrhythmias and CAD,

leaving a research gap in other heart diseases and

medical tasks, such as treatment.

(MQ6): Physiological signals and Electronic

Health Records (EHR) were the main data types,

highlighting the underutilization of other data types.

(MQ7): Binary classification, the dominant

approach, was often linked to the screening task.

(MQ8): Traditional ML techniques were

predominantly used in most studies, suggesting the

need for researchers to investigate more innovative

techniques for classification in cardiology.

REFERENCES

Almazroi, A. A., Aldhahri, E. A., Bashir, S., & Ashfaq, S.

(2023). A Clinical Decision Support System for Heart

Disease Prediction Using Deep Learning. IEEE Access,

99, 1-1, https://doi.org/10.1109/ACCESS.2023.328524.

Amazal, F. A., & Idri, A. (2019). Handling of categorical

data in software development effort estimation: A

systematic mapping study. In Proceedings of the 2019

Federated Conference on Computer Science and

Information Systems, FedCSIS 2019 (pp. 763–770).

https://doi.org/10.15439/2019F222.

Araki, T., et al. (2016, March). PCA-based Polling Strategy

in Machine Learning Framework for Coronary Artery

Disease Risk Assessment in Intravascular Ultrasound:

A Link between Carotid and Coronary Grayscale

Plaque Morphology. Computer Methods and Programs

in Biomedicine, 128(8). https://doi.org/10.1016/j.cmpb.2

016.02.004.

Aziz, F., et al. (2021, August). Short- and long-term

mortality prediction after an acute ST-elevation

myocardial infarction (STEMI) in Asians: A machine

learning approach. PLoS ONE, 16(8), e0254894.

https://doi.org/10.1371/journal.pone.0254894.

Chicco, D., & Jurman, G. (2020, February). Machine

learning can predict survival of patients with heart

failure from serum creatinine and ejection fraction

alone. BMC Medical Informatics and Decision Making,

20(1). https://doi.org/10.1186/s12911-020-1023-5.

Collet, J.-P., et al. (2020, August). 2020 ESC Guidelines for

the management of acute coronary syndromes in

patients presenting without persistent ST-segment

elevation. European Heart Journal, 42(14).

https://doi.org/10.1093/eurheartj/ehaa575.

Dangare, C. S., & Apte, S. S. (2012, June). Improved study

of heart disease prediction system using data mining

classification techniques. International Journal of

Computer Applications, 47(10), 44-48.

https://doi.org/10.5120/7228-0076.

Dwivedi, A. K. (2018, May). Performance evaluation of

different machine learning techniques for prediction of

Machine Learning Classiﬁcation in Cardiology: A Systematic Mapping Study

415

heart disease. Neural Computing and Applications.

https://doi.org/10.1007/s00521-016-2604-1.

Esfandiari, N., et al. (2014, July). Knowledge discovery in

medicine: Current issue and future trend. Expert

Systems with Applications, 41(9), 4434–4463.

https://doi.org/10.1016/j.eswa.2014.01.011.

Garg, P., Steen, J., & Mardon, A. A. (2022, February). An

exhaustive comparative analytical study of 15 machine

learning models for automated cardiovascular disease

classification. Paper presented at the 2021 4th

International Conference on Recent Trends in

Computer Science and Technology (ICRTCST)

https://doi.org/10.1109/ICRTCST54752.2022.9781828.

Guo, Y., et al. (2023, May). Machine learning-enhanced

echocardiography for screening coronary artery

disease. BioMedical Engineering OnLine, 22(1).

https://doi.org/10.1186/s12938-023-01106-x.

Hassan, C. A. U., et al. (2022, September). Effectively

predicting the presence of coronary heart disease using

machine learning classifiers. Sensors, 22(19), 7227.

https://doi.org/10.3390/s22197227.

Noh, K., Lee, H. G., Shon, H.-S., Lee, B. J., & Ryu, K. H.

(2006, October). Associative classification approach

for diagnosing cardiovascular disease. Intelligent

Computing in Signal Processing and Pattern

Recognition (pp. 721-727). Lecture Notes in Control

and Information Sciences, 345. https://doi.org/10.1007/

978-3-540-37258-5_82.

Sun, D.-W. (2008, January). Computer Vision Technology

for Food Quality Evaluation. https://doi.org/10.1016/

B978-0-12-373642-0.X5001-7.

Rahman, Q. A., et al. (2015, April). Utilizing ECG-Based

Heartbeat Classification for Hypertrophic

Cardiomyopathy Identification. IEEE Transactions on

Nanobioscience, 14(5). https://doi.org/10.1109/TNB.20

15.2426213.

Idri, A., Benhar, H., Fernández-Alemán, J. L., & Kadi, I.

(2018, May). A systematic map of medical data

preprocessing in knowledge discovery. Computer

Methods and Programs in Biomedicine, 162.

https://doi.org/10.1016/j.cmpb.2018.05.007.

Idri, A., Amazal, F. A., & Abran, A. (2014, August).

Analogy-based software development effort

estimation: A systematic mapping and review.

Information and Software Technology, 58.

https://doi.org/10.1016/j.infsof.2014.07.013.

Juhola, M., Joutsijoki, H., Penttinen, K., & Aalto-Setälä, K.

(2018, June). Detection of genetic cardiac diseases by

Ca2+ transient profiles using machine learning

methods. Scientific Reports, 8(1). https://doi.org/10.10

38/s41598-018-27695-5.

Kadi, I., Idri, A., & Fernandez-Aleman, J. L. (2017,

January). Knowledge discovery in cardiology: A

systematic literature review. International Journal of

Medical Informatics, 97, 12-32. https://doi.org/10.10

16/j.ijmedinf.2016.09.005.

Kadi, I., Idri, A., & Fernandez-Aleman, J. L. (2019,

September). Systematic mapping study of data mining–

based empirical studies in cardiology. Health

Informatics Journal, 25(14), 146045821771763.

https://doi.org/10.1177/1460458217717636.

Kitchenham, B., Budgen, D., & Brereton, P. (2010,

January). The value of mapping studies: A participant-

observer case study.

Kitchenham, B., Pearl, O., Budgen, D., Turner, M., Bailey,

J., & Linkman, S. (2009, January). Systematic literature

reviews in software engineering: A systematic literature

review. Information and Software Technology, 51(1),

7-15. https://doi.org/10.1016/j.infsof.2008.09.009.

Puyol Anton, E., et al. (2021). Statistical Atlases and

Computational Models of the Heart. M&Ms and

EMIDEC Challenges 11th International Workshop,

STACOM 2020, Held in Conjunction with MICCAI

2020, Lima, Peru, October 4, 2020, Revised Selected

Papers. Lecture Notes in Computer Science. Springer.

https://doi.org/10.1007/978-3-030-68107-4.

Seetharam, K., et al. (2022). Applications of Machine

Learning in Cardiology. Cardiology and Therapy,

11(24). https://doi.org/10.1007/s40119-022-00273-7.

Smole, T., et al. (2021). A machine learning-based risk

stratification model for ventricular tachycardia and

heart failure in hypertrophic cardiomyopathy.

Computers in Biology and Medicine, 135, 104648.

https://doi.org/10.1016/j.compbiomed.2021.104648.

Swathy, M., & Saruladha, K. (2021). A comparative study

of classification and prediction of Cardio-Vascular

Diseases (CVD) using Machine Learning and Deep

Learning techniques. ICT Express, 8(3).

https://doi.org/10.1016/j.icte.2021.08.021.

DATA 2024 - 13th International Conference on Data Science, Technology and Applications

416