A Supplier Selection Model Based on P-SVM with GA

In this study, a potential support vector machines (P- SVM) with genetic algorithm (GA) is proposed to supplier selection. P-SVM is used to select optimal classifier using "support feature" by exchanging the roles of data points and features. On the other hand, one-against-one method is applied to solve the problem of multi-class. In addition, genetic algorithm is applied to accomplish the appropriate parameters selection so as to improve the performance of P-SVM as much as possible. The results of simulations show that the generalization performance of the methods based on P-SVM is higher than the ones based on standard SVM and can accomplish more scientific and reasonable criteria definition through training of P-SVM.