Improved adaptive neuro-fuzzy inference system based on modified glowworm swarm and differential evolution...

Area of Science:

• Computational Intelligence
• Medical Informatics
• Machine Learning

Background:

• Modern hybrid techniques are advancing medical diagnosis for disease screening and management.
• Accurate prediction of neurodegenerative (e.g., glaucoma, Parkinson's) and carcinogenic (e.g., breast cancer) diseases remains a critical challenge.

Purpose of the Study:

• To develop and evaluate a novel predictive model, DE-GSO-ANFIS, for early detection of complex diseases.
• To enhance the efficiency of the Adaptive Neuro-Fuzzy Inference System (ANFIS) through a hybrid optimization approach.

Main Methods:

• The proposed model integrates Differential Evolution (DE) with Glowworm Swarm Optimization (GSO) to improve ANFIS parameter estimation.
• The DE-GSO algorithm addresses the local minima problem inherent in the standard GSO algorithm.
• Performance is benchmarked against traditional ANFIS, GA-ANFIS, PSO-ANFIS, LOA-ANFIS, DE-ANFIS, and GSO.

Main Results:

• The DE-GSO-ANFIS model demonstrated significantly improved predictive accuracy for medical disorders.
• Experimental results confirmed the superiority of the proposed hybrid approach over existing methods.
• The enhanced optimization effectively improved ANFIS performance in disease prediction tasks.

Conclusions:

• The DE-GSO-ANFIS model offers a robust and efficient solution for predicting neurodegenerative and carcinogenic diseases.
• Hybrid optimization strategies show great promise for advancing medical diagnostic tools.
• This approach can aid in earlier disease detection and improved patient management.