Comparative analysis of machine learning models to predict rectangular patch antenna dimensions

The non-linear relationship between antenna characteristics and their electromagnetic responses made the design and optimization process difficult, in the face of these difficulties, antenna engineers used several techniques, including machine learning, as it has great capabilities that make it a very useful tool that can help researchers in this field. In this paper, four machine learning algorithms: ANN, Random Forest, Decision Tree and SVR are used to predict the dimensions of a rectangular patch antenna, through the utilization of a dataset comprising 3111 simulated samples collected using HFSS Simulation software. The results showed that random forest with 100 estimators exhibited an outstanding performance concerning prediction accuracy, with a mean square error (MSE) of 0.52.

• rectangular patch antenna
• machine learning
• random forest
• artificial neural networks
• decision tree
• support vector regression
• antenna optimization

• Antenna Design and Analysis
• Antenna Design and Optimization
• Machine Learning in Materials Science
• rectangular patch antenna
• machine learning
• random forest
• artificial neural networks
• decision tree
• support vector regression
• antenna optimization
• Aerospace Engineering
• Materials Chemistry
• Engineering
• Materials Science