The Modified P-SVM for Regional Logistics System Evaluation

To evaluate the regional logistics,the usual method is to make the model by fixation index and expert weightiness,which cannot show the complex relationship.The artificial intelligence method was applied to solve it.The decision tree was applied to the P-SVM method in order to improve the evaluation result,which keeps the original advantages and the class capacity at the same time.The index on evaluation of regional logistics system was put forward based on reference review firstly.Moreover,a modified approach of support vector machine,potential support vector machine,combined with decision tree is introduced based on analysing the problems faced by evaluation of regional logistics system.The simulation results show that the model based on modified potential support vector machine not only has much stronger generalization ability but also has the ability of feature selection which is not owned by other ones based on standard support vector machine.