Chaotic simulated annealing multi-verse optimization enhanced kernel extreme learning machine for medical diagnosis

Area of Science:

• Computational intelligence
• Medical informatics
• Optimization algorithms

Background:

• Multi-Verse Optimization (MVO) is crucial for disease diagnosis but suffers from slow convergence and local optima.
• Existing MVO methods struggle with high-dimensional and multi-modal problems, limiting their diagnostic accuracy.

Purpose of the Study:

• To enhance the efficiency and accuracy of Multi-Verse Optimization (MVO) for disease diagnosis.
• To propose a novel hybrid model, CSAMVO-KELM, by integrating an improved MVO with Kernel Extreme Learning Machine (KELM).

Main Methods:

• A new Chaotic Simulated Annealing overhaul of MVO (CSAMVO) was developed, incorporating a chaotic intensification mechanism (CIP) and simulated annealing (SA).
• The CSAMVO algorithm was benchmarked against classical algorithms on 39 functions to assess solution quality and convergence speed.
• A hybrid CSAMVO-KELM model was established and evaluated on two disease diagnosis problems against competitive classifiers.

Main Results:

• CSAMVO demonstrated superior performance over other algorithms in terms of solution quality and convergence speed.
• The CSAMVO-KELM model achieved higher predictive performance and better learning potential in disease diagnosis tasks.
• The proposed CSAMVO approach effectively mitigates MVO's limitations, including slow convergence and local optima entrapment.

Conclusions:

• The developed CSAMVO algorithm offers a more stable and efficient convergence, overcoming the limitations of traditional MVO.
• The CSAMVO-KELM hybrid model presents a promising advancement for accurate and efficient disease diagnosis.
• This research highlights the potential of enhanced optimization techniques in improving medical diagnostic systems.