An attribute weight assignment and particle swarm optimization algorithm for medical database classifications.

Area of Science:

• Computational intelligence
• Medical informatics
• Machine learning

Background:

• Accurate medical data classification is crucial for effective diagnosis and treatment.
• Existing methods may face challenges with complex medical datasets.
• Hybrid models offer potential for improved classification performance.

Purpose of the Study:

• To develop and evaluate a novel hybrid model for medical data classification.
• To integrate Case-Based Reasoning (CBR) and Particle Swarm Optimization (PSO) for enhanced diagnostic accuracy.
• To assess the model's performance on benchmark medical datasets.

Main Methods:

• A hybrid Case-Based Reasoning Particle Swarm Optimization (CBRPSO) model was developed.
• CBR was used for data preprocessing and deriving feature weights.
• PSO was employed to construct a decision-making system, optimizing cluster reduction for classification.

Main Results:

• The CBRPSO model achieved an average forecasting accuracy of 97.4% for breast cancer diagnosis.
• The model achieved an average forecasting accuracy of 76.8% for liver disorder diagnosis.
• The hybrid approach demonstrated improved accuracy and comprehensibility for medical experts.

Conclusions:

• The proposed CBRPSO model is effective for medical data classification.
• This hybrid approach offers a promising tool for medical diagnosis.
• The model provides accurate and interpretable results, aiding medical professionals.