Novel Feature Selection and Voting Classifier Algorithms for COVID-19 Classification in CT Images.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Accurate COVID-19 diagnosis is crucial, yet challenging due to overlapping symptoms with other pneumonias.
• Chest CT scans are vital for diagnosis but processing them is computationally intensive.
• Machine learning offers a promising solution for efficient and accurate COVID-19 detection from medical images.

Purpose of the Study:

• To propose novel optimization algorithms for feature selection and classification of COVID-19 from CT scans.
• To develop an efficient framework for automated COVID-19 diagnosis, reducing computational cost.
• To enhance the accuracy of COVID-19 detection using advanced machine learning techniques.

Main Methods:

• Features were extracted using AlexNet (Convolutional Neural Network).
• A Guided Whale Optimization Algorithm (Guided WOA) with Stochastic Fractal Search (SFS) was used for feature selection and balancing.
• A Guided WOA voting classifier integrated with Particle Swarm Optimization (PSO) aggregated predictions from SVM, NN, KNN, and DT classifiers.

Main Results:

• The proposed SFS-Guided WOA feature selection algorithm demonstrated superior efficiency compared to existing methods.
• The PSO-Guided-WOA voting classifier achieved an Area Under the Curve (AUC) of 0.995, outperforming other voting classifiers.
• Statistical tests (Wilcoxon, ANOVA, T-test) confirmed the high quality and statistical significance of the proposed algorithms.

Conclusions:

• The developed framework provides an accurate and computationally efficient method for COVID-19 diagnosis using CT images.
• The novel optimization algorithms (SFS-Guided WOA and PSO-Guided-WOA) show significant potential for medical image analysis.
• This approach can aid clinicians in faster and more reliable COVID-19 diagnosis, improving patient outcomes.