Quantum-Behaved Particle Swarm Optimization for Parameter Optimization of Support Vector Machine

Abstract

Support vector machine (SVM) parameters such as penalty parameter and kernel parameters have a great influence on the complexity and accuracy of SVM model. In this paper, quantum-behaved particle swarm optimization (QPSO) has been employed to optimize the parameters of SVM, so that the classification error can be reduced. To evaluate the proposed model (QPSO-SVM), the experiment adopted seven standard classification datasets which are obtained from UCI machine learning data repository. For verification, the results of the QPSO-SVM algorithm are compared with the standard PSO, and genetic algorithm (GA) which is one of the well-known optimization algorithms. Moreover, the results of QPSO are compared with the grid search, which is a conventional method of searching parameter values. The experimental results demonstrated that the proposed model is capable to find the optimal values of the SVM parameters. The results also showed lower classification error rates compared with standard PSO and GA algorithms.