Enhancing Heart Disease Classification Accuracy and Computational Efficiency Using Machine Learning and Feature Optimization

Despite recent advances in the field of diagnostics and treatment, heart disease is one of the leading causes of death in the world and it is so important to have an early and accurate diagnosis. This paper investigates the use of machine learning classifiers and optimization techniques towards improving prediction of heart disease. The dataset of 14 features was used. First, machine learning classifiers such as Naïve Bayes (NB), k-Nearest Neighbors (kNN) and Decision Tree (DT) were applied to the full dataset. These classifiers were then measured for their performance in terms of accuracy, precision, recall and F1 score. In order to further optimize the results, the most effective features were selected based on three optimization techniques Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and Ant Colony Optimization (ACO) to reduce redundancy and computational complexity. The classifiers were then applied again to the optimized datasets, and it showed significant improvements in their predictive efficiency. kNN consistently performed well at all datasets with accuracy improvement from 71.72% to 98.53% after feature optimization. This work shows how optimization improves accuracy, and also the computational efficiency. This study emphasizes the ability to use optimized machine learning models as a suitable and inexpensive means for clinical decision making in order to predict heart disease.