Liver Cancer Algorithm: A novel bio-inspired optimizer.

Area of Science:

• Computational Intelligence
• Bio-inspired Computing
• Optimization Algorithms

Background:

• Optimization algorithms are crucial for solving complex computational problems.
• Existing algorithms often struggle with balancing local and global search capabilities.
• Bio-inspired approaches offer novel strategies for enhancing optimization performance.

Purpose of the Study:

• To introduce a novel bio-inspired optimization algorithm, the Liver Cancer Algorithm (LCA).
• To evaluate the LCA's performance on standard benchmark functions and compare it with existing metaheuristic algorithms.
• To demonstrate the LCA's applicability to real-world problems, specifically feature selection in biomedical data classification.

Main Methods:

• The Liver Cancer Algorithm (LCA) was developed, simulating liver tumor growth and takeover dynamics.
• LCA incorporates genetic operators and a Random Opposition-Based Learning (ROBL) strategy for effective search space exploration.
• The algorithm was benchmarked against seven established metaheuristic algorithms on the CEC'2020 test suite.

Main Results:

• The LCA algorithm demonstrated competitive performance against established algorithms on CEC'2020 benchmark functions.
• The LCA-SVM model, combining LCA with Support Vector Machines, achieved high accuracy in biomedical data classification.
• Specifically, the LCA-SVM model attained 98.704% accuracy on the MonoAmine Oxidase (MAO) dataset.

Conclusions:

• The Liver Cancer Algorithm (LCA) is an effective bio-inspired optimization technique.
• LCA shows superior performance in both mathematical optimization and practical feature selection tasks.
• The LCA-SVM model offers a promising approach for accurate biomedical data classification.