A novel approach for heart disease prediction using hybridized AITH²O algorithm and SANFIS classifier

Area of Science:

• Cardiology
• Artificial Intelligence
• Machine Learning

Background:

• Heart disease poses a significant global health threat, leading to high mortality and morbidity.
• Accurate early prediction of heart disease is crucial for effective patient treatment and medical recommendations.
• Existing machine learning models face challenges with computational complexity and overfitting, limiting diagnostic accuracy.

Purpose of the Study:

• To develop a novel hybrid optimization algorithm for improved heart disease classification accuracy.
• To optimize computation time in smart healthcare applications for heart disease prediction.
• To address the limitations of existing machine learning classifiers in terms of accuracy and efficiency.

Main Methods:

• Feature selection using the hybrid Arithmetic Optimization and Inter-Twinned Mutation-Based Harris Hawk Optimization (AITH²O) algorithm.
• Parameter tuning of the Stabilized Adaptive Neuro-Fuzzy Inference System (SANFIS) classifier with the AITH²O algorithm.
• Validation using the Cleveland heart disease dataset and simulation in MATLAB 2020a.

Main Results:

• The proposed hybrid AITH²O algorithm demonstrates strong exploration and exploitation capabilities with faster convergence.
• The hybrid SANFIS classifier achieved a superior accuracy of 99.28% for heart disease prediction.
• A true positive rate of 99.46% was attained, outperforming existing state-of-the-art techniques.

Conclusions:

• The developed hybrid optimization algorithm (AITH²O) effectively enhances the accuracy and efficiency of heart disease prediction.
• The Stabilized Adaptive Neuro-Fuzzy Inference System (SANFIS) classifier, optimized by AITH²O, offers a promising solution for smart healthcare diagnostics.
• This approach provides a robust and accurate method for early detection and management of heart disease.