Using support vector machine in credit scorecard development

Abstract

Support vector machines, came into alive from statistical learning theory, means a new stage in the evoution of datamining algorithms. The method handle the problem of classification in a very natural way, and promise growth is the efficiency of learning statistical patterns. However in Basel II required parameter estimation and in credit scorecard development support vector machines are not wide range used. In this paper it will be shown how to use SVMs for calculating probability of default, placing emphasis on business understanding and impact on expected loss calculation.