Hybrid genetic algorithms-driven optimization of machine learning models for heart disease prediction

Area of Science:

• Cardiology
• Computer Science
• Bioinformatics

Background:

• Machine learning models like KNN and SVM are crucial for heart disease prediction.
• Hyperparameter selection significantly impacts the performance of these ML models.
• Existing methods often struggle with optimal hyperparameter tuning.

Purpose of the Study:

• To introduce a novel hybrid approach using Genetic Algorithm (GA) for optimizing KNN and SVM hyperparameters.
• To enhance the classification accuracy and clinical relevance of ML models for heart disease prediction.
• To systematically improve the predictive performance of machine learning models.

Main Methods:

• Utilized a Genetic Algorithm (GA) to systematically optimize hyperparameters for K-Nearest Neighbour (KNN) and Support Vector Machine (SVM) classifiers.
• Evaluated model performance based on key classification metrics including accuracy, precision, recall, and F-score.
• Applied GA-driven hyperparameter tuning to improve predictive outcomes.

Main Results:

• GA-based hyperparameter tuning led to substantial performance improvements in both KNN and SVM models.
• Achieved a classification accuracy of 95.38% for KNN and 90% for SVM.
• Demonstrated significant enhancements in precision, recall, and F-score for the optimized models.

Conclusions:

• GA-based hyperparameter tuning is an effective strategy for boosting the performance of ML models in heart disease classification.
• The proposed hybrid approach offers improved predictive power and clinical relevance for heart disease diagnosis.
• Optimized ML models show potential for more accurate and reliable heart disease prediction.