Research on a new information fusion method based on SVM and SCQPSO and its application

The optimal parameters of the support vector machine (SVM) are very important for accuracy modeling and generalization performance. The quantum particle swarm optimization (QPSO) algorithm takes on the characteristics of the rapid global optimization, scale chaos method provides the characteristics of the fast convergence and the SVM has the characteristics of the nonlinear fitting. These advantages of the scale chaos method and the QPSO algorithm are used to propose a scale chaos QPSO (SCQPSO) algorithm. Then the SCQPSO algorithm is used to optimize the parameters of the SVM model. A new information fusion method based the SCQPSO algorithm and the SVM model (SCQPSO-SVM) is proposed in this paper. The SCQPSO-SVM algorithm uses the global optimization ability of the SCQPSO algorithm to comprehensively optimize the penalty coefficient, kernel parameter and hybrid weight of the SVM model. The goal is to improve the solved speed and solution accuracy of the SVM model. The SCQPSO-SVM algorithm is applied in the testing function and the rotor fault diagnosis of traction motor. The experimental results show that the SCQPSO algorithm can search for the good optimization results and the SCQPSO-SVM algorithm can reduce the error rate of the fusion recognition. So the SCQPSO-SVM algorithm takes on better generalization performance and prediction accuracy in the real application.