A Survey of Ensemble Learning: Concepts, Algorithms, Applications, and Prospects

IEEE Access, Journal Year: 2022, Volume and Issue: 10, P. 99129 - 99149

Published: Jan. 1, 2022

Ensemble learning techniques have achieved state-of-the-art performance in diverse machine applications by combining the predictions from two or more base models. This paper presents a concise overview of ensemble learning, covering three main methods: bagging, boosting, and stacking, their early development to recent algorithms. The study focuses on widely used algorithms, including random forest, adaptive boosting (AdaBoost), gradient extreme (XGBoost), light (LightGBM), categorical (CatBoost). An attempt is made concisely cover mathematical algorithmic representations, which lacking existing literature would be beneficial researchers practitioners.

Language: Английский

---

Towards Diagnostic Aided Systems in Coronary Artery Disease Detection: A Comprehensive Multiview Survey of the State of the Art

International Journal of Intelligent Systems, Journal Year: 2023, Volume and Issue: 2023, P. 1 - 19

Published: Aug. 9, 2023

Introduction. Coronary artery disease (CAD) is one of the main causes death all over world. One way to reduce mortality rate from CAD predict its risk and take effective interventions. The use machine learning- (ML-) based methods an method for predicting CAD-induced death, which why many studies in this field have been conducted recent years. Thus, study aimed review published on artificial intelligence classification algorithms detection diagnosis. Methods. This systematically reviewed most cutting-edge techniques analyzing clinical paraclinical data quickly diagnose CAD. We searched PubMed, Scopus, Web Science databases using a combination related keywords. A extraction form was used collect after selecting articles inclusion exclusion criteria. content analysis analyze data, study’s objectives, results are presented tables figures. Results. Our search three prevalent resulted 15689 studies, 54 were included be analysis. Most laboratory demographic shown desirable results. In general, ML (traditional ML, DL/NN, ensemble) used. Among used, random forest (RF), linear regression (LR), neural networks (NNs), support vector (SVM), K-nearest (KNNs) applications code recognition. Conclusion. findings show that these models different successful despite lack benchmark comparing features, methods, Many performed better their analyses features as result closer look. near future, specialists can ML-based powerful tool diagnosing more precisely by looking at design’s technical facets. incredible outcomes decreased diagnostic errors, time, needless invasive tests, typically decreases expenses healthcare systems.

Language: Английский

---

RF-CNN-F: random forest with convolutional neural network features for coronary artery disease diagnosis based on cardiac magnetic resonance

Scientific Reports, Journal Year: 2022, Volume and Issue: 12(1)

Published: July 1, 2022

Coronary artery disease (CAD) is a prevalent with high morbidity and mortality rates. Invasive coronary angiography the reference standard for diagnosing CAD but costly associated risks. Noninvasive imaging like cardiac magnetic resonance (CMR) facilitates assessment can serve as gatekeeper to downstream invasive testing. Machine learning methods are increasingly applied automated interpretation of other clinical results medical diagnosis. In this study, we proposed novel detection method based on CMR images by utilizing feature extraction ability deep neural networks combining features aid random forest very first time. It necessary convert image data numeric so that they be used in nodes decision trees. To end, predictions multiple stand-alone convolutional (CNNs) were considered input The capability CNNs representing renders our generic classification approach applicable any dataset. We named RF-CNN-F, which stands Random Forest CNN Features. conducted experiments large dataset have collected made publicly accessible. Our achieved excellent accuracy (99.18%) using Adam optimizer compared trained fivefold cross validation (93.92%) tested same

Language: Английский

---

A Powerful Paradigm for Cardiovascular Risk Stratification Using Multiclass, Multi-Label, and Ensemble-Based Machine Learning Paradigms: A Narrative Review

Diagnostics, Journal Year: 2022, Volume and Issue: 12(3), P. 722 - 722

Published: March 16, 2022

Cardiovascular disease (CVD) causes the highest mortality globally. With escalating healthcare costs, early non-invasive CVD risk assessment is vital. Conventional methods have shown poor performance compared to more recent and fast-evolving Artificial Intelligence (AI) methods. The proposed study reviews three most paradigms for assessment, namely multiclass, multi-label, ensemble-based in (i) office-based (ii) stress-test laboratories.

Language: Английский

---

Solving the class imbalance problem using ensemble algorithm: application of screening for aortic dissection

BMC Medical Informatics and Decision Making, Journal Year: 2022, Volume and Issue: 22(1)

Published: March 28, 2022

Abstract Background Imbalance between positive and negative outcomes, a so-called class imbalance, is problem generally found in medical data. Despite various studies, imbalance has always been difficult issue. The main objective of this study was to find an effective integrated approach address the problems posed by validate method early screening model for rare cardiovascular disease aortic dissection (AD). Methods Different data-level methods, cost-sensitive learning, bagging were combined solve low sensitivity caused two classes First, feature selection applied select most relevant features using statistical analysis, including significance test logistic regression. Then, we assigned different misclassification cost values classes, constructed weak classifiers based on support vector machine (SVM) model, with undersampling methods build final strong classifier. Due rarity AD, data particularly prominent. Therefore, our construction AD disease. Clinical 523,213 patients from Institute Hypertension, Xiangya Hospital, Central South University used verify validity method. In these data, sample ratio non-AD 1:65, each contained 71 features. Results proposed ensemble achieved highest 82.8%, training time specificity reaching 56.4 s 71.9% respectively. Additionally, it obtained small variance 19.58 × 10 –3 seven-fold cross validation experiment. results outperformed common algorithms AdaBoost, EasyEnsemble, Random Forest (RF) as well single learning (ML) regression, decision tree, k nearest neighbors (KNN), back propagation neural network (BP) SVM. Among five ML algorithms, SVM after performed best 79.5% 73.4%. Conclusions study, demonstrate that integration selection, undersampling, can overcome challenge dataset develop practical which could lead at stage.

Language: Английский

---

Artificial intelligence in cardiovascular prevention: new ways will open new doors

Journal of Cardiovascular Medicine, Journal Year: 2023, Volume and Issue: 24(Supplement 2), P. e106 - e115

Published: May 1, 2023

Prevention and effective treatment of cardiovascular disease are progressive issues that grow in tandem with the average age world population. Over recent decades, potential role artificial intelligence medicine has been increasingly recognized because incredible amount real-world data (RWD) regarding patient health status healthcare delivery can be collated from a variety sources wherein information is routinely collected, including registries, clinical case reports, reimbursement claims billing medical devices, electronic records. Like any other (health) data, RWD analysed accordance high-quality research methods, its analysis deliver valuable patient-centric insights complementing obtained conventional trials. Artificial application on to detect patient's trajectory leading personalized tailored treatment. This article reviews benefits prevention management, focusing diagnostic therapeutic improvements without neglecting limitations this new scientific approach.

Language: Английский

---

Revolutionizing Cardiology through Artificial Intelligence—Big Data from Proactive Prevention to Precise Diagnostics and Cutting-Edge Treatment—A Comprehensive Review of the Past 5 Years

Diagnostics, Journal Year: 2024, Volume and Issue: 14(11), P. 1103 - 1103

Published: May 26, 2024

Background: Artificial intelligence (AI) can radically change almost every aspect of the human experience. In medical field, there are numerous applications AI and subsequently, in a relatively short time, significant progress has been made. Cardiology is not immune to this trend, fact being supported by exponential increase number publications which algorithms play an important role data analysis, pattern discovery, identification anomalies, therapeutic decision making. Furthermore, with technological development, have appeared new models machine learning (ML) deep (DP) that capable exploring various cardiology, including areas such as prevention, cardiovascular imaging, electrophysiology, interventional many others. sense, present article aims provide general vision current state use cardiology. Results: We identified included subset 200 papers directly relevant research covering wide range applications. Thus, paper presents arithmology, clinical or emergency procedures summarized manner. Recent studies from highly scientific literature demonstrate feasibility advantages using different branches Conclusions: The integration cardiology offers promising perspectives for increasing accuracy decreasing error rate efficiency practice. From predicting risk sudden death ability respond cardiac resynchronization therapy diagnosis pulmonary embolism early detection valvular diseases, shown their potential mitigate feasible solutions. At same limits imposed small samples studied highlighted alongside challenges presented ethical implementation; these relate legal implications regarding responsibility making processes, ensuring patient confidentiality security. All constitute future directions will allow

Language: Английский

---

Clinical Application of Artificial Intelligence in the Diagnosis, Prediction, and Classification of Coronary Heart Disease

Cardiovascular Innovations and Applications, Journal Year: 2025, Volume and Issue: 10(1)

Published: Jan. 1, 2025

Coronary heart disease (CHD), the most common cause of mortality globally, poses a formidable challenge to modern healthcare systems. Artificial intelligence (AI) is playing an increasingly important role in multiple diagnostic applications CHD, by facilitating early intervention and personalized treatment. This mini review describing state art provides clinicians with updated insights into transformative potential AI enhance CHD detection. can be used increase prognostic accuracy. However, its reliance on homogeneous numerical data might potentially lead misdiagnoses unnecessary radiation exposure diagnosing CHD. Multimodal fusion brings new for accurate diagnosis medicine. Finally, unmet challenges future research directions ethical, regulatory, technical aspects are discussed. aimed at bridging gap between advancements practical clinical settings, achieve which empowers context ecosystem.

Language: Английский

---

Machine Learning Predictive Models for Coronary Artery Disease

SN Computer Science, Journal Year: 2021, Volume and Issue: 2(5)

Published: June 22, 2021

Language: Английский

---

Diagnostic Classification and Prognostic Prediction Using Common Genetic Variants in Autism Spectrum Disorder: Genotype-Based Deep Learning

JMIR Medical Informatics, Journal Year: 2021, Volume and Issue: 9(4), P. e24754 - e24754

Published: March 14, 2021

In the United States, about 3 million people have autism spectrum disorder (ASD), and around 1 out of 59 children are diagnosed with ASD. People ASD characteristic social communication deficits repetitive behaviors. The causes this remain unknown; however, in up to 25% cases, a genetic cause can be identified. Detecting as early possible is desirable because detection enables timely interventions Identification based on objective pathogenic mutation screening major first step toward intervention effective treatment affected children. Recent investigation interrogated genomics data for detecting treating disorders, addition conventional clinical interview diagnostic test. Since deep neural networks perform better than shallow machine learning models complex high-dimensional data, study, we sought apply obtained across thousands simplex families at risk identify contributory mutations create an advanced classifier screening. After preprocessing from Simons Simplex Collection, extracted top ranking common variants that may protective or chi-square A convolutional network-based was then designed using identified significant predict autism. performance compared learning-based classifiers randomly selected variants. were significantly enriched chromosome X while Y also discriminatory determining identification autistic individuals nonautistic individuals. ARSD, MAGEB16, MXRA5 genes had largest effect Thus, algorithms adapted include these model yielded area under receiver operating curve 0.955 accuracy 88% identifying Our demonstrated considerable improvement ~13% terms classification standard tools. Common informative identification. findings suggest process reliable method distinguishing diseased group control

Language: Английский

---