Detecting COVID-19 patients based on fuzzy inference engine and Deep Neural Network

Area of Science:

• Medical Diagnostics
• Computational Biology
• Infectious Disease Research

Background:

• Coronavirus disease (COVID-19) poses a significant global health threat, necessitating rapid and accurate detection methods.
• Current diagnostic approaches for COVID-19 have limitations in detection efficiency.
• Timely identification of infected individuals is crucial for controlling the spread of COVID-19.

Purpose of the Study:

• To introduce a novel Hybrid Diagnose Strategy (HDS) for the accurate detection of COVID-19.
• To develop an improved feature selection and classification approach for COVID-19 diagnosis.
• To enhance the efficiency and reliability of COVID-19 diagnostic tools.

Main Methods:

• Developed a Hybrid Diagnose Strategy (HDS) involving a novel feature ranking technique within a proposed Patient Space (PS).
• Constructed a Feature Connectivity Graph (FCG) to determine feature weights and interdependencies.
• Utilized a hybrid classification model combining a fuzzy inference engine and a Deep Neural Network (DNN) for patient classification.

Main Results:

• The proposed HDS achieved high performance metrics: 97.658% accuracy, 96.756% precision, 96.55% recall, and 96.615% F-measure.
• HDS demonstrated a low error rate of 2.342%, outperforming existing techniques.
• Statistical validation using Wilcoxon Signed Rank Test and Friedman Test confirmed the robustness of the results.

Conclusions:

• The Hybrid Diagnose Strategy (HDS) offers a highly accurate and efficient method for COVID-19 detection.
• HDS shows significant potential in improving diagnostic capabilities for infectious diseases like COVID-19.
• The developed approach provides a reliable tool for identifying infected individuals, aiding public health efforts.