Enhanced prediction of metamaterial antenna parameters using advanced machine learning regression models

Al Gburi, Ahmed Jamal Abdullah and Jain, Prince and Sahoo, Prabodh Kumar and Khaleel, Aymen Dheyaa (2024) Enhanced prediction of metamaterial antenna parameters using advanced machine learning regression models. Progress In Electromagnetics Research C, 146. pp. 1-12. ISSN 1937-8718

[img] Text
02702260820241820531053.pdf
Download (5MB)

Abstract

The integration of machine learning (ML) regression models in predicting the parameters of metamaterial antennas significantly reduces the design time required for optimizing antenna performance compared to traditional simulation tools. Metamaterial antennas, known for overcoming the bandwidth constraints of small antennas, benefit greatly from these advanced predictive models. This study applies and evaluates four ML regression models — Extra Trees, Random Forest, XGBoost, and CatBoost — to predict key antenna parameters such as S11, gain, and bandwidth. Each model’s performance is assessed using metrics like Mean Absolute Error (MAE), MeanSquaredError (MSE), R-squared (R2), Mean Absolute Percentage Error (MAPE), and Root Mean Squared Error (RMSE) across different training and testing set configurations (30%, 50%, and 70%). The Extra Trees model achieves the best performance for predicting gain, with an R2 of 0.9990, MAE of 0.0069, MSE of 0.0002, RMSE of 0.0145, and MAPE of 0.3106. Feature importance analysis reveals that specific features, such as pr and p0 for gain and Y a and Xa for bandwidth, are critical in the predictive models. These findings highlight the potential of ML methods to improve the efficiency and accuracy of metamaterial antenna design

Item Type: Article
Uncontrolled Keywords: Metamaterial Antenna, Machine Learning, Regression Models, Parameter Prediction, Electromagnetic Simulation
Divisions: Faculty Of Electronics And Computer Technology And Engineering
Depositing User: Norfaradilla Idayu Ab. Ghafar
Date Deposited: 06 Jan 2025 10:01
Last Modified: 06 Jan 2025 10:01
URI: http://eprints.utem.edu.my/id/eprint/28123
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item