Adaptive Memetic Binary Optimization (AMBO) Algorithm

A novel adaptive memetic binary optimization algorithm for feature selection

AMBO: Adaptive Memetic Binary Optimization Algorithm for Feature Selection

This repository contains the official MATLAB implementation of the AMBO (Adaptive Memetic Binary Optimization) algorithm proposed in the paper:

A. C. Çınar, A novel adaptive memetic binary optimization algorithm for feature selection, Artificial Intelligence Review, 2023. DOI: 10.1007/s10462-023-10482-8

📌 About the Project

AMBO is a pure binary metaheuristic algorithm specifically designed for feature selection tasks. It uses:

• Adaptive crossover mechanisms (single-point, double-point, uniform)
• Canonical mutation
• Logic gate-based local search using AND, OR, and XOR for balancing exploration and exploitation.

It has been tested on 21 benchmark datasets and outperformed several state-of-the-art algorithms including BPSO, GA variants, BDA, BSSA, and BGWO.

📂 Files

• Main.m: Main script to run the algorithm.
• datasets/: Sample datasets used in the paper.
• results/: Contains output logs and performance results.

🧪 Requirements

• MATLAB R2021a or later
• Statistics and Machine Learning Toolbox (for KNN)

📈 Citation

If you use this code or data in your research, please cite the paper as:

@article{cinar2023ambo,

title={A novel adaptive memetic binary optimization algorithm for feature selection},

author={Cinar, Ahmet Cevahir},

journal={Artificial Intelligence Review},

year={2023},

doi={10.1007/s10462-023-10482-8}

}

🤝 Collaboration

Contributions, ideas, and collaborations are welcome!

Feel free to contact me for research partnerships, extensions, or comparative benchmarking:

📧 Email: [e-mail]

🔗 LinkedIn: Ahmet Cevahir Çınar