2024_Enhancement Of Heart Disease Prediction Performance with a Hybrid Feature Selection Technique And Stacking Ensemble Method

Bibliographic Details
Format: General Document
_version_ 1860798296106205184
building INTELEK Repository
collection Online Access
collectionurl https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3
copyright Copyright©PWB2025
country Malaysia
date 2024-10-25 09:43
format General Document
id 16935
institution UniSZA
originalfilename 16935_7edf41c55e632a5.pdf
recordtype oai_dc
resourceurl https://intelek.unisza.edu.my/intelek/pages/view.php?ref=16935
sourcemedia Server storage
Scanned document
spelling 16935 https://intelek.unisza.edu.my/intelek/pages/view.php?ref=16935 https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 General Document Malaysia Library Staff (Top Management) Library Staff (Management) Library Staff (Support) Terengganu Faculty of Informatics & Computing English application/pdf 1.7 Microsoft® Word for Microsoft 365 124 Server storage Scanned document UniSZA Private Access UniSZA Copyright©PWB2025 Machine Learning Dissertations, Academic Ensemble learning (Machine learning) Medical informatics Statistical methods Heart Disease Prediction Stacking Ensemble Methods UniSZA Heart Diseases Diagnosis Data processing 2024_Enhancement Of Heart Disease Prediction Performance with a Hybrid Feature Selection Technique And Stacking Ensemble Method Recent years have witnessed a transformative shift in the field of heart disease prediction, driven by the widespread adoption of advanced machine learning and data analytics techniques. Heart disease remains a common global health dilemma, with substantial implications for both morbidity and mortality across the world. The fact that nearly 12 million lives are lost yearly emphasizes the central challenge facing healthcare professionals. In most countries, a shortage of cardiovascular specialists endures, accompanied by a troubling rate of incorrectly diagnosed cases. To tackle these challenges, it is crucial to establish a precise and efficient early-stage prediction model for detecting heart diseases in patients. This is further supported by the utilization of analytical instruments aimed at enhancing the effectiveness of the clinical decision-making process. Presently, previous research tends to lack comprehensive investigations, as most research leans towards using single classifiers and may not encompass the full spectrum of available features in their experiments leading to lower accuracy for the prediction of heart disease.In addressing this issue, the research aims to explore the effectiveness of an ensemble stacking classification method in combination with a feature selection technique. This research tends to bridge this gap by employing the UCI dataset comprising 303 records and 13 attributes, aiming to explore the efficacy of the selected method. Specifically, a hybrid approach feature selection method was employed to combine chi-squared and analysis of variance (ANOVA), aimed to filter out the most crucial features for subsequent evaluation with chosen classifiers. Ten base classifiers are assessed including logistic regression (LR), random forest (RF), k-nearest neighbor (KNN), decision tree (DT), naïve bayes (NB), support vector classifier (SVC), extra gradient boosting (XGB), multi-layer perceptron (MLP), stochastic gradient descent (SGD), and extra tree classifier (ETC). In comparison, the meta-model classifiers consisting of LR, SVC, MLP, and NB each displaying different levels of accuracy are tested with the selected features from chi-squared and ANOVA.As a result, 93.44% accuracy achieved by naïve bayes stands out as the most promising outcome, surpassing individual classifiers. The result distinctly showcases the strength of ensemble techniques and hybrid feature selection in enhancing the accuracy of heart disease prediction. In conclusion, this study investigates the efficiency of combining an ensemble stacking method with a feature selection technique to predict heart disease. By using a hybrid feature selection approach that incorporates chi-squared and ANOVA, essential features are effectively pinpointed for subsequent evaluation with chosen classifiers. This research sets the stage for the creation of more precise and dependable heart disease prediction tools, with the potential to enhance clinical decision-making and ultimately benefit patient outcomes. 2024-10-25 09:43 uuid:8390477B-758C-4769-B602-C03D6497ED43 16935_7edf41c55e632a5.pdf ANOVA Ensemble Stacking Chi-Squared Test Artificial Intelligence in Healthcare Clinical Decision Support Systems UCI Dataset Heart — Diseases — Diagnosis — Data processing Artificial intelligence — Medical applications Thesis
spellingShingle 2024_Enhancement Of Heart Disease Prediction Performance with a Hybrid Feature Selection Technique And Stacking Ensemble Method
state Terengganu
subject Dissertations, Academic
Ensemble learning (Machine learning)
Medical informatics Statistical methods
Heart Diseases Diagnosis Data processing
Heart — Diseases — Diagnosis — Data processing
Artificial intelligence — Medical applications
summary Recent years have witnessed a transformative shift in the field of heart disease prediction, driven by the widespread adoption of advanced machine learning and data analytics techniques. Heart disease remains a common global health dilemma, with substantial implications for both morbidity and mortality across the world. The fact that nearly 12 million lives are lost yearly emphasizes the central challenge facing healthcare professionals. In most countries, a shortage of cardiovascular specialists endures, accompanied by a troubling rate of incorrectly diagnosed cases. To tackle these challenges, it is crucial to establish a precise and efficient early-stage prediction model for detecting heart diseases in patients. This is further supported by the utilization of analytical instruments aimed at enhancing the effectiveness of the clinical decision-making process. Presently, previous research tends to lack comprehensive investigations, as most research leans towards using single classifiers and may not encompass the full spectrum of available features in their experiments leading to lower accuracy for the prediction of heart disease.In addressing this issue, the research aims to explore the effectiveness of an ensemble stacking classification method in combination with a feature selection technique. This research tends to bridge this gap by employing the UCI dataset comprising 303 records and 13 attributes, aiming to explore the efficacy of the selected method. Specifically, a hybrid approach feature selection method was employed to combine chi-squared and analysis of variance (ANOVA), aimed to filter out the most crucial features for subsequent evaluation with chosen classifiers. Ten base classifiers are assessed including logistic regression (LR), random forest (RF), k-nearest neighbor (KNN), decision tree (DT), naïve bayes (NB), support vector classifier (SVC), extra gradient boosting (XGB), multi-layer perceptron (MLP), stochastic gradient descent (SGD), and extra tree classifier (ETC). In comparison, the meta-model classifiers consisting of LR, SVC, MLP, and NB each displaying different levels of accuracy are tested with the selected features from chi-squared and ANOVA.As a result, 93.44% accuracy achieved by naïve bayes stands out as the most promising outcome, surpassing individual classifiers. The result distinctly showcases the strength of ensemble techniques and hybrid feature selection in enhancing the accuracy of heart disease prediction. In conclusion, this study investigates the efficiency of combining an ensemble stacking method with a feature selection technique to predict heart disease. By using a hybrid feature selection approach that incorporates chi-squared and ANOVA, essential features are effectively pinpointed for subsequent evaluation with chosen classifiers. This research sets the stage for the creation of more precise and dependable heart disease prediction tools, with the potential to enhance clinical decision-making and ultimately benefit patient outcomes.
title 2024_Enhancement Of Heart Disease Prediction Performance with a Hybrid Feature Selection Technique And Stacking Ensemble Method
title_full 2024_Enhancement Of Heart Disease Prediction Performance with a Hybrid Feature Selection Technique And Stacking Ensemble Method
title_fullStr 2024_Enhancement Of Heart Disease Prediction Performance with a Hybrid Feature Selection Technique And Stacking Ensemble Method
title_full_unstemmed 2024_Enhancement Of Heart Disease Prediction Performance with a Hybrid Feature Selection Technique And Stacking Ensemble Method
title_short 2024_Enhancement Of Heart Disease Prediction Performance with a Hybrid Feature Selection Technique And Stacking Ensemble Method
title_sort 2024_enhancement of heart disease prediction performance with a hybrid feature selection technique and stacking ensemble method