Computational Intelligence-Based Disease Severity Identification: A Review of Multidisciplinary Domains.

Area of Science:

• Computational intelligence and artificial intelligence in healthcare.
• Machine learning for disease diagnosis and severity assessment.

Background:

• Disease severity identification is crucial for effective treatment planning.
• Computational intelligence (CI) and deep learning (DL) offer promising solutions for accurate and rapid disease severity assessment.
• The multi-class classification nature of disease severity necessitates advanced computational methods.

Purpose of the Study:

• To provide a comprehensive review of recent computational intelligence-based approaches for disease severity identification.
• To analyze methodologies, datasets, and performance metrics of reviewed studies.
• To highlight key research trends and public data repositories for disease severity identification.

Main Methods:

• Systematic literature review following PRISMA guidelines.
• Compilation of studies from the last decade focusing on multidisciplinary diseases.
• Analysis of research papers based on methodology, datasets, performance metrics (accuracy, specificity), and disease types.

Main Results:

• Identified numerous CI-based solutions for disease severity identification across various disciplines.
• Detailed examination of approaches for Parkinson's Disease and Diabetic Retinopathy severity.
• Brief coverage of severity identification for other conditions including COVID-19, Alzheimer's, and various cancers.

Conclusions:

• CI approaches are increasingly vital for accurate disease severity identification.
• The review serves as a compendium and offers insights for future research in the field.
• Public repositories are available to facilitate further research in computational intelligence for disease severity.