Machine Learning Techniques for C-slot Loaded Minkowski Fractal Antenna

Abstract For Wi-MAX wireless applications, a C-slot-loaded Minkowski fractal microstrip antenna is examined and its operating frequency is predicted using machine learning techniques. A Minkowski fractal antenna is first created by substituting Minkowski curves along the square patch's sides. A C-slot is inserted afterwards. For parametric analysis, the minkowski fractal curve’s indentation depth, the length of the C-slot arm, and the width are changed. The training data set for the proposed C-slot-loaded minkowski fractal antenna resonance frequency is developed by altering the indentation depth, C-slot length, and width. The training data set is then evaluated using machine learning methods such as multivariate regression, ANN, XGBoost, random forest, and decision tree. XGBoost machine learning outscored them all and had a very low mean square error.