A hybrid bio inspired neural model based on Ropalidia Marginata behavior for multi disease classification.

Area of Science:

• Biomedical data analysis
• Machine learning in healthcare
• Computational intelligence

Background:

• Accurate disease diagnosis is crucial but challenging.
• Machine learning, particularly artificial neural networks (ANNs), shows promise for intelligent disease detection.
• Optimizing ANNs is key to improving classification accuracy and avoiding local minima.

Purpose of the Study:

• To present a novel hybrid Ropalidia Marginata Optimization-based hybrid neural network (RMO-NN) for enhanced medical data classification.
• To leverage biologically inspired mechanisms for optimizing neural network learning.
• To reduce classification errors in disease detection.

Main Methods:

• Developed a hybrid RMO-NN integrating Ropalidia Marginata Optimization (RMO) with neural networks.
• Incorporated RMO's task allocation and dominance hierarchy for neural network optimization.
• Validated RMO-NN on breast cancer, diabetes, blood transfusion, and medical image datasets.

Main Results:

• RMO-NN demonstrated superior performance over Cuckoo Search Neural Network (CSNN) and Artificial Bee Colony Neural Network (ABCNN) in accuracy, MSE, SD, and convergence speed.
• The model showed significant improvements on biomedical data classification tasks.
• RMO-NN outperformed state-of-the-art deep learning models on medical image datasets.

Conclusions:

• The proposed RMO-NN is an effective approach for improving disease diagnosis through enhanced medical data classification.
• The novel RMO algorithm offers significant advantages in optimizing neural network performance.
• This work contributes a powerful tool for accurate and efficient biomedical data analysis.