Combining market and accounting-based models for credit scoring using a classification scheme based on support vector machines

•Combination of option-based model with accounting data for credit risk model.•Application of market model to non-listed firms.•Use of a novel additive support vector machines model.

Credit risk rating is an important issue for both financial institutions and companies, especially in periods of economic recession. There are many different approaches and methods which have been developed over the years. The aim of this paper is to create a credit risk rating model, using a machine learning methodology that combines accounting data with the option-based approach of Black, Scholes, and Merton. The model is built on data for companies listed in the Greek stock exchange, but it is also shown to provide accurate results for non-listed firms as well. Linear and nonlinear support vector machines are used for model building, as well as an innovative additive modeling approach, which enables the construction of comprehensible and accurate credit scoring models.