Scholarlink

<p>This approach optimizes data processing for enhancing efficiency and accuracy of predictive models in diagnosing heart disease. To determine the most effective models for this, various machine learning (ML) algorithms are applied to dataset. A metaheuristic feature selection algorithm is employed to refine the selection of relevant features, thereby improving predictive performance, reducing computational complexity, and enhancing generalization. The goal of this study is to develop a robust ML-based model to improve the diagnosis of heart disease and forecast patient survival outcomes, contributing to better healthcare decision-making. This study focuses on creating predictive models using Logistic Regression (LR), K-Nearest Neighbor Classifier (KNNC), Support Vector Machine (SVM), Naïve Bayes Classifier (NBC), Decision Tree Classifier (DTC), Random Forest Classifier (RFC), and Gradient Boost Classifier (GBC). Gradient Boosting Classifier (GBC) achieved the highest accuracy of 91%, surpassing other machine learning algorithms in performance. To further improve predictive accuracy and efficiency, metaheuristic feature selection algorithms including Cuckoo Search (CS), Flower Pollination Algorithm (FPA), Whale Optimization Algorithm (WOA), and Harris Hawks Optimization (HHO) were integrated. These advanced optimization techniques refine feature selection enhancing model’s ability to diagnose heart disease with greater precision. We developed a methodology for identifying Leveraging Machine Learning Techniques: Heart Disease Prediction researchers in medicine believe that forecasting is essential for potential patients with heart disease. However, selecting most representative features for medical research can be difficult. Features were chosen from Cleveland dataset utilizing CS, FPA, WOA, and HHO. A prediction model for heart disease was developed using machine learning techniques (MLT), specifically LR, KNNC, SVM, NBC, DTC, RFC, and GBC. An 80:20 split ratio for training and testing data was used for evaluation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Heart Disease Prediction Using Machine Learning with Metaheuristic Feature Selection Approaches

  • 期刊论文

  • 发表日期：

    2025年10月15日

查看全文
  • Salliah Shafi,
  • Gufran Ahmad Ansari

摘要

This approach optimizes data processing for enhancing efficiency and accuracy of predictive models in diagnosing heart disease. To determine the most effective models for this, various machine learning (ML) algorithms are applied to dataset. A metaheuristic feature selection algorithm is employed to refine the selection of relevant features, thereby improving predictive performance, reducing computational complexity, and enhancing generalization. The goal of this study is to develop a robust ML-based model to improve the diagnosis of heart disease and forecast patient survival outcomes, contributing to better healthcare decision-making. This study focuses on creating predictive models using Logistic Regression (LR), K-Nearest Neighbor Classifier (KNNC), Support Vector Machine (SVM), Naïve Bayes Classifier (NBC), Decision Tree Classifier (DTC), Random Forest Classifier (RFC), and Gradient Boost Classifier (GBC). Gradient Boosting Classifier (GBC) achieved the highest accuracy of 91%, surpassing other machine learning algorithms in performance. To further improve predictive accuracy and efficiency, metaheuristic feature selection algorithms including Cuckoo Search (CS), Flower Pollination Algorithm (FPA), Whale Optimization Algorithm (WOA), and Harris Hawks Optimization (HHO) were integrated. These advanced optimization techniques refine feature selection enhancing model’s ability to diagnose heart disease with greater precision. We developed a methodology for identifying Leveraging Machine Learning Techniques: Heart Disease Prediction researchers in medicine believe that forecasting is essential for potential patients with heart disease. However, selecting most representative features for medical research can be difficult. Features were chosen from Cleveland dataset utilizing CS, FPA, WOA, and HHO. A prediction model for heart disease was developed using machine learning techniques (MLT), specifically LR, KNNC, SVM, NBC, DTC, RFC, and GBC. An 80:20 split ratio for training and testing data was used for evaluation.