Optimization of SVM Parameters using Gaussian Quantum-Behaved PSO

This research seeks to offer insights regarding optimizing support vector machines (SVM) in terms of their parameters which are critical for streamlining accurate classifications for different tasks. We introduce a novel optimization approach called Gaussian quantum-behaved particle swarm optimization (GQPSO), denoted GQPSO-SVM, that is meant to minimize the test error rate very efficiently by finding the optimal SVM parameters. A number of experiments were conducted to compare the GQPSO-SVM algorithm with wellknown techniques including BA+SVM, PSO+SVM, and QPSO+SVM. The results show that GQPSO-SVM regularly achieves lower test error rates than its alternatives with the radial basis function (RBF) kernel performing better than the polynomial kernel. This illustrates how GQPSO-SVM can improve SVM classifier performance and dependability across a range of applications.