Artificial plant optimization algorithm to detect heart rate & presence of heart disease using machine learning

Area of Science:

• Biomedical Engineering
• Machine Learning
• Cardiovascular Health

Background:

• Cardiovascular diseases, particularly Coronary Heart Disease (CHD), are a leading cause of mortality worldwide.
• Timely prediction of CHD is crucial for effective intervention and treatment.
• Existing methods require efficient feature selection for accurate prediction.

Purpose of the Study:

• To introduce a Modified Artificial Plant Optimization (MAPO) algorithm for optimal feature selection.
• To predict heart rate using fingertip video datasets.
• To enhance the prediction accuracy of Coronary Heart Disease (CHD) using machine learning models.

Main Methods:

• Pre-processing and noise filtering of fingertip video datasets.
• Application of MAPO for heart rate prediction and feature optimization.
• Integration of predicted heart rate as a feature in subsequent datasets for CHD prediction using various machine learning algorithms.

Main Results:

• MAPO achieved a high Pearson correlation (0.9541) and low Standard Error Estimate (2.418) in heart rate prediction.
• MAPO demonstrated significant dimensionality reduction (up to 81.25%) in datasets.
• Machine learning models utilizing MAPO-optimized features showed comparable accuracies in CHD prediction.

Conclusions:

• MAPO is a highly effective feature selection algorithm for cardiovascular disease prediction.
• The proposed method offers a promising approach for non-invasive CHD detection through fingertip video analysis.
• MAPO outperforms other optimization techniques in terms of accuracy and dimensionality reduction.