GA-ANFIS Expert System Prototype for Prediction of Dermatological Diseases

Area of Science:

• Artificial Intelligence
• Medical Informatics
• Dermatology

Background:

• Dermatological disease detection relies on accurate feature analysis.
• Existing methods may lack optimal feature selection and model optimization.
• A robust expert system is needed for real-world dermatological diagnoses.

Purpose of the Study:

• To develop and validate a novel expert system prototype, GA-ANFIS, for dermatological disease detection.
• To optimize fuzzy inference systems using Genetic Algorithms for improved diagnostic accuracy.
• To analyze the impact of dermatological features on disease detection.

Main Methods:

• Utilized nine dermatological features as inputs for classification.
• Employed Adaptive Neuro-Fuzzy Inference Systems (ANFIS) for initial fuzzy model optimization.
• Applied Genetic Algorithms (GA) for a second level of fuzzy model optimization within the GA-ANFIS system.
• Performed system modeling and validation using MATLAB and a validation dataset.

Main Results:

• The GA-ANFIS system demonstrated higher accuracy rates in dermatological disease detection compared to the standard ANFIS model.
• Analysis provided insights into the influence of specific features on diagnostic outcomes.
• The proposed GA-ANFIS model proved effective in a real-world data context.

Conclusions:

• The GA-ANFIS expert system offers a significant advancement in automated dermatological disease detection.
• Integrating Genetic Algorithms with ANFIS enhances the performance of fuzzy inference systems for medical diagnoses.
• The developed system shows promise for clinical application in dermatology.