Supervised Machine Learning Model for Accurate Output Prediction of Various Antenna Designs

In this work, a detailed study on the performance of a machine learning algorithm (k-nearest neighbors) in the prediction of antenna output response of three different antenna designs is carried out. Each of the selected design have different number of design parameters. The effect of training input data set size of independent and identically distributed (IID’s) antenna design variables on S11 response prediction has been analyzed using root mean squared error (RMSE) of each design. The KNeighborsRegressor tool available in scikit-learn package within Python is employed to train and test the KNN model.