Credit Default Prediction - Data-Driven Comprehensive Approach for Credit Default Prediction of Non-Listed Austrian Companies

This research focuses on the largely overlooked credit risk assessment of non-listed companies, particularly SMEs. While prior studies mainly address large, listed firms, we utilize an extensive dataset from KSV1870 containing over 1.7 million balance sheets since 1995. By integrating financial data with macroeconomic and other non-financial information, we adopt a holistic approach to default prediction. The study employs both traditional methods, such as logistic regression, and modern techniques such as random forests and support vector machines. Our goal is to enhance data-driven risk assessment for non-listed firms, supporting creditors in their decision-making processes.