A new parallel multi-objective Harris hawk algorithm for predicting the mortality of COVID-19 patients

Area of Science:

• Computational intelligence
• Bio-inspired algorithms
• Medical informatics

Background:

• Predicting COVID-19 patient mortality is crucial for resource allocation and treatment strategies.
• Feature selection is vital for building efficient and accurate predictive models in healthcare.

Purpose of the Study:

• To develop a novel parallel multi-objective Harris' Hawk Optimization (HHO) algorithm.
• To predict COVID-19 patient mortality risk using symptom data.
• To optimize feature selection for enhanced prediction accuracy and reduced dimensionality.

Main Methods:

• Implementation of a parallel multi-objective HHO algorithm.
• Application to a real-world COVID-19 dataset, including an augmented version.
• Comparison with existing state-of-the-art metaheuristic wrapper algorithms.

Main Results:

• Achieved 98.15% prediction accuracy with a 45% reduction in features.
• Demonstrated significant improvements over existing methods.
• Showcased the effectiveness of feature selection in improving classification performance.

Conclusions:

• The proposed parallel multi-objective HHO algorithm is effective for COVID-19 mortality risk prediction.
• Feature selection enhances predictive model performance and efficiency.
• This approach offers a promising tool for clinical decision support in managing COVID-19 patients.