The Ability of Support Vector Machine (SVM) in Financial Distress Prediction

Predicting financial distress, which normally happens before bankruptcy, is a challenging phenomenon and a crucial issue in all firms. The importance of data mining tools is well recognized, such that nowadays they are widely used in different financial issues such as, prediction of bankruptcy, financial distress, company's performance prediction, management fraud discovery and credit risk assessment. Using support vector machine and combinations of cash flow components, this research attempts to predict financial distress of companies. Combinations of cash flows, as input variables (data) of the model, are selected based on specific criteria of financial distress. Results reveal that among Kernel functions of the model, polynomial function has the most power of prediction in year of financial distress or one and two years prior to year of distress.