Diseases diagnosis based on artificial intelligence and ensemble classification.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Machine Learning Applications

Background:

• Computer Aided Diagnosis (CAD) is crucial for enhancing medical diagnostic systems.
• Integrating Artificial Intelligence (AI) into CAD systems offers significant potential for automation and error reduction.
• Traditional manual workflows in medical diagnosis are often slow, inaccurate, and prone to human error.

Purpose of the Study:

• To present a comprehensive Computer Aided Disease Diagnosis (CAD²) strategy utilizing Machine Learning (ML).
• To assist clinicians in making more informed and accurate medical decisions through an enhanced diagnostic approach.

Main Methods:

• The proposed CAD² strategy comprises three sequential phases: Outlier Rejection Phase (ORP), Feature Selection Phase (FSP), and Classification Phase (CP).
• ORP employs a novel Outlier Rejection Technique (ORT) with Fast Outlier Rejection (FOR) and Accurate Outlier Rejection (AOR) stages.
• FSP utilizes a Hybrid Selection Technique (HST) combining fisher score (QS²) and a Hybrid Bio-inspired Optimization (HBO) technique (PS²). CP relies on Ensemble Classification Technique (ECT).

Main Results:

• Experimental validation on two datasets (general diseases and COVID-19) demonstrated the high efficiency of the CAD² strategy.
• The proposed CAD² system outperformed existing diagnostic strategies in terms of accuracy, error rates, precision, and recall.
• Statistical tests (Wilcoxon signed rank and Friedman) confirmed the superior performance of CAD².

Conclusions:

• The CAD² strategy, incorporating ORP, FSP, and CP, provides accurate disease diagnosis.
• It achieves higher accuracy, lower error rates, and reduced implementation time compared to alternative diagnostic methods.