Paper Title
Comparison of Artificial Intelligence Models to Predict Phases of Patch Array Antenna

Abstract
The present study introduces some artificial intelligence (AI) model for the lobe direction calculation of an antenna. Each model utilizes the direction of the lobe as input and the phases of antenna element as outputs. Thus, the lobe direction of the antenna is readily derived from the trained model's output, which is the phases of the antenna element. While, it’s not easy to derive the phases of each antenna element from the desired lobe direction. For this purpose, Decision tree Regressor, chain SVM regressor and direct SVM regressor are three AI models used for simulation. 1280 lobe direction record for 8x1 array patch antenna were generated by changing the phases of the antenna element, and this set of data were utilized to train each model. The performance of the trained models is tested by estimating the Mean Absolute Error percentage (MAPE) for each. The obtained results are compared and direct SVM regressor model have the lowest MAPE value which provides its efficiency in lobe direction prediction. Keywords - Artificial Intelligence, Lobe Direction, Microstrip patch array antenna, Machine Learning.