Combining Pairwise SVM Classifiers for Bond Rating

Prior studies on financial applications of data mining techniques mainly focused on the applicability of artificial neural networks, but recent research tended to investigate applicability of novel data mining techniques just like support vector machines. Support vector machines (SVMs) were originally devised for binary classification. But, in reality, there exist many problems which cannot be solved by just binary classification models, those are multiclass classification problems including bond rating. Researchers have tried to extend original SVM to multiclass classification. However, their studies have only focused on classifying samples into nominal categories. This study proposes Ordinal Multiclass SVMs which apply ordinal pairwise partitioning (OPP) to conventional SVMs in order to handle ordinal multiple classes efficiently and effectively. Our suggested model may use fewer classifiers but predict more accurately because it utilizes additional hidden information, the order of the classes. To validate our model, we apply it to the real-world bond rating case. In this study, we compare the results of the model to those of other multiclass classification algorithms. The result shows that Ordinal Multiclass SVMs improve prediction performance.