Rule Extraction for Screening of COVID-19 Disease Using Granular Computing Approach

Area of Science:

• Medical Informatics
• Computational Intelligence
• Epidemiology

Background:

• Inadequate access to treatment centers is a major challenge during viral epidemics like Coronavirus.
• Premature visits to medical facilities for mild symptoms contribute to virus spread and healthcare congestion.
• Early symptom detection is crucial for timely diagnosis and effective disease management.

Purpose of the Study:

• To present a novel approach for patient classification and disease diagnosis using granular computing.
• To address the challenge of early detection for unknown viral infections, specifically Coronavirus.
• To improve the efficiency of healthcare services by enabling accurate, symptom-based diagnosis.

Main Methods:

• Utilizing granular computing for patient classification and disease diagnosis based on symptoms.
• Employing expert-classified training datasets for rule extraction.
• Focusing on the extraction of rules with zero entropy for high accuracy.

Main Results:

• The proposed granular computing method successfully detects Coronavirus disease (COVID-19) based on observed symptoms.
• The approach facilitates accurate classification of patients, aiding in early diagnosis.
• Experimental results validate the effectiveness of the zero-entropy rule extraction.

Conclusions:

• Granular computing offers a powerful framework for symptom-based disease diagnosis.
• The developed method can aid in early detection of viral infections, mitigating epidemic spread.
• This approach has the potential to optimize healthcare resource allocation during public health crises.