Support Vector Machine (SVM) employs Structural Risk Minimization (SRM) principle to generalize better than conventional machine learning methods employing the traditional Empirical Risk Minimization (ERM) principle. When applying SVM to response modeling in direct marketing, however, one has to deal with the practical difficulties: large training data, class imbalance and binary SVM output. This paper proposes ways to alleviate or solve the addressed difficulties through informative sampling, use of different costs for different classes, and use of distance to decision boundary. This paper also provides various evaluation measures for response models in terms of accuracies, lift chart analysis and computational efficiency.

Best Paper Award