Regression Analysis of Metamaterial Antenna using Decision and Extra Tree Regressors

Designing an antenna is a task that is both complicated and time-consuming, demanding a thorough comprehension of electromagnetic theory and the physics of antennas. Conventionally, designers rely on analytical models and simulations to evaluate the antenna's performance, which can be computationally expensive and time intensive. However, machine learning offers a potential solution to automate and optimize the antenna design process by leveraging data-driven techniques to identify the best design parameters and configurations. The paper proposes the use of machine learning algorithms using hyper parameter tuning for regression to predict various antenna parameters, such as Return Loss, Voltage Standing Wave Ratio, and Gain. Based on the prediction results, an R squared of 0.9998 was obtained for VSWR, whereas for Gain an R squared of 0.9638. For Return loss an R squared of 0.9950 was achieved. The proposed regression system could be adopted in an antenna design unit to test the simulations and predict the antenna performance.