Hybrid strategy enhanced crayfish optimization algorithm for breast cancer prediction.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Machine Learning

Background:

• The standard Crayfish Optimization Algorithm (COA) faces challenges including reduced diversity, limited exploration, and premature convergence to local optima.
• These limitations hinder its effectiveness in complex optimization problems and real-world applications.

Purpose of the Study:

• To introduce a hybrid strategy enhanced crayfish optimization algorithm (MSCOA) that addresses the limitations of the traditional COA.
• To improve global exploration, local search effectiveness, and overall population evolution efficiency.

Main Methods:

• Implemented a chaotic inverse exploration initialization for enhanced population diversity and global exploration.
• Utilized an adaptive t-distributed feeding strategy to increase population variety and local search effectiveness.
• Introduced an adaptive ternary optimization mechanism with acceleration factors and dynamic weight adjustment for refined search intensity.

Main Results:

• MSCOA demonstrated superior convergence accuracy and robustness compared to traditional COA on CEC2005 and CEC2019 benchmark datasets.
• Statistical analysis (Wilcoxon test, p < 0.05) confirmed MSCOA's superiority over five other algorithms.
• The MSCOA-ELM model achieved 100% accuracy and F1 score on the Wisconsin breast cancer dataset, a significant improvement over baseline ELM.

Conclusions:

• The proposed MSCOA effectively enhances diversity, exploration, and convergence accuracy in optimization problems.
• MSCOA shows strong potential for practical applications, evidenced by its performance in medical data classification.
• Further research is recommended to explore additional improvements and applications of the MSCOA algorithm.