Binary Information Press
To solve multi-category classification task and asymmetric misclassification costs in customer loyalty segmentation, a multi-category support vector machine is proposed. The support vector machines for multi-classification integrate with a set of binary classification support vector machine and use cost sensitive function to minimize the asymmetric misclassification cost. The empirical data including attributes from customer satisfaction survey and credit card transaction history is used to validate the proposed model. The results show that the proposed multi-category support vector machine can enhance the accuracy of customer loyalty classification as well as distinguish high value customer better.