A Novel COVID-19 Diagnostic System Using Biosensor Incorporated Artificial Intelligence Technique.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Infectious Disease Modeling

Background:

• COVID-19 poses a significant global health threat with high incidence and mortality.
• Accurate and timely diagnosis of COVID-19 disease state, types, and recovery is crucial for effective patient management and public health strategies.
• Existing diagnostic methods may have limitations in speed, accuracy, or ability to categorize disease progression.

Purpose of the Study:

• To introduce and evaluate a novel deep learning model, the Shuffle Shepherd Optimization-based Generalized Deep Convolutional Fuzzy Network (SSO-GDCFN), for COVID-19 diagnosis.
• To assess the capability of SSO-GDCFN in classifying COVID-19 disease states, types, and recovered categories.
• To demonstrate the superior performance of SSO-GDCFN compared to traditional diagnostic techniques.

Main Methods:

• Development of a hybrid deep learning architecture integrating fuzzy logic and convolutional neural networks.
• Optimization of the network using the Shuffle Shepherd Optimization algorithm.
• Training and validation of the SSO-GDCFN model on a dataset for COVID-19 diagnosis and classification.

Main Results:

• The proposed SSO-GDCFN achieved exceptional diagnostic accuracy of 99.99%.
• Key performance metrics include 99.98% precision, 100% sensitivity/recall, 95% specificity, 0.965% kappa, 0.88% AUC, and Mean Squared Error (MSE) below 0.07%.
• The model demonstrated rapid processing with results obtained in approximately 25 seconds, outperforming conventional methods in accuracy and minimizing reclassification errors.

Conclusions:

• The SSO-GDCFN model presents a highly accurate and efficient tool for diagnosing and categorizing COVID-19.
• This advanced computational approach offers significant potential for improving clinical decision-making in infectious disease management.
• The study highlights the effectiveness of integrating optimization algorithms with deep fuzzy networks for complex medical diagnostic tasks.