Integration of Particle Swarm Optimization (PSO) and Machine Learning to Improve Classification Accuracy During Antenna Design

Flexibility and generalization of antenna analysis and synthesis using artificial neural networks have attracted enormous attention in the field of microwave-strip antennas (MSAs). Various problems can be solved using Particle Swarm Optimization PSO by applying social connections. A Swarm of particles (agents) is applied to find the best possible solution. These agents seek the space coordinates associated with the best solution each particle has so far achieved. Predicting response times with a trained neural network is almost equivalent to measuring or simulating them. The proposed work has integrated an optimization mechanism into machine learning which improves the reliability of the model. The simulation shows the categorization process utilizing machine learning for antenna design. As a result of using the PSO optimizer, the accuracy, precision, f1-score, and recall value of the simulations have significantly improved. An improvement of 4%, 10%, and 5% is observed in accuracy, recall value, and in precision by using the optimization technique. The result proves that the integration of machine learning into the optimizer increases the suggested model's dependability.