Optimization of Machine Learning Algorithms with Bagging and AdaBoost Methods for Stroke Disease Prediction

Aug 2, 2026

Upload by Patrick

Stroke is an acute neurologic disorder caused by impaired blood flow to the brain, leading to insufficient oxygen and remaining one of the leading global causes of death. Machine learning can support health professionals by enabling earlier stroke prediction. This study compares stroke classification performance using Bagging and AdaBoost with Naïve Bayes, SVM, Decision Tree, and KNN on the Kaggle Stroke Prediction dataset. Decision Tree achieves the highest baseline accuracy at 91%, and ensemble optimization improves only this algorithm, with modest gains across accuracy, precision, recall, and F1-score depending on missing-value handling.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Abstract

Introduction

Material and Method

Optimization of Machine Learning Algorithms with Bagging and AdaBoost Methods for Stroke Disease Prediction Helmi Saifullah MANSUR, Nelly Oktavia ADIWIJAYA*, Tio DHARMAWAN Department of Computer Science, University of Jember, Jalan Kalimantan No. 37 Jember, East Java, 68121, Indonesia. [E-mail: helmimansur88@gmail.com](E-mail: helmimansur88@gmail.com); (*) [nelly.oa@unej.ac.id](nelly.oa@unej.ac.id); [tio.pssi@unej.ac.id](tio.pssi@unej.ac.id) * Author to whom correspondence should be addressed; Received: May 25, 2023 /Accepted: June 25, 2023/ Published online: July 1, 2023 Abstract Stroke is an acute neurologic disorder of blood vessels in the brain due to blockage of blood flow to the brain resulting in less oxygen. Stroke remains one of the leading causes of death worldwide. Therefore, developing Machine Learning is expected to help health professionals make early predictions of stroke. This study aimed to compare the performance results of stroke classification modeling using Bagging and AdaBoost methods in Machine Learning algorithms (Naïve Bayes, Support Vector Machine, Decision Tree, and K-Nearest Neighbors) using Stroke Prediction Dataset from Kaggle. The results show that Machine Learning algorithm that has the best performance is Decision Tree with 91% accuracy, followed by KNN, Naïve Bayes, and finally, SVM. Optimization of Machine Learning algorithms with Bagging and AdaBoost only increases the performance value of the Decision Tree algorithm but does not increase the performance value of other algorithms. The results of Decision Tree optimization with Bagging increased 1% accuracy and F1-score, as well as 4% precision in the missing value deleted scenario. Furthermore, in the missing value scenario using mean value increases 1% F1-score and 4% precision. While the results of Decision Tree optimization with AdaBoost increased 1% recall and F1-score in the missing value deleted scenario. Then in the missing value scenario using mean value has the same performance as without optimization. The application of Bagging and AdaBoost methods only increases the performance value of the Decision Tree algorithm, but the increase is still insignificant. Keywords: Stroke; Machine Learning (ML); Bagging; AdaBoost Introduction Stroke is an acute neurological dysfunction of the blood vessels in the brain caused by the cessation of blood supply to the brain so that brain cells lack the necessary oxygen [1] . Based on the 2019 Global Burden of Disease (GBD) information, stroke remains the second leading cause of death and the third leading cause of combined death and disability in the world [2] . Basic Health Research data in Indonesia stated that in 2013 was 12.1 per mile of national stroke prevalence, while in 2018, it reported a prevalence of 10.9 per mile population, with the highest values in East Kalimantan Province (14.7 per mile) and the lowest in Papua Province (4.1 per mile) [3] . Based on its type, stroke can be ischemic stroke or hemorrhagic stroke [4] . Ischemic stroke occurs due to blockage of blood vessels by a thrombus or embolus, resulting in brain ischemia, while hemorrhagic stroke occurs due to bleeding and rupture of weakened blood vessels around the brain tissue, causing intracranial pressure [5] . With the development of information and communication technology in both the fields of Artificial Intelligence (AI) and Machine Learning (ML), it is hoped that it can take an important role in making early predictions to treat various diseases, one of which is stroke [1] . Machine learning techniques have been widely used in multiple healthcare applications in recent years [6] . Machine learning can be a useful step towards efficient treatment in early stroke detection and assist healthcare professionals in making clinical decisions and predictions. Research in the last few decades, machine learning has been used in improving stroke diagnosis in terms of accuracy and speed [5] . Sailasya and Kuma

FAQs

Which machine learning algorithm achieved the best baseline performance for stroke prediction?

The Decision Tree algorithm achieved the best baseline performance with 91% accuracy, followed by KNN, Naïve Bayes, and SVM.

How do Bagging and AdaBoost affect the performance of different algorithms?

Bagging and AdaBoost optimization increases performance only for the Decision Tree algorithm, while the other algorithms do not show performance improvements.

What impact does missing-value handling have on Decision Tree optimization results?

With Bagging, missing value deleted increases accuracy and F1-score and raises precision; using mean value increases F1-score and precision. With AdaBoost, missing value deleted increases recall and F1-score, while mean value results match the baseline performance.

Aug 2, 2026

Upload by Patrick