Comparing Mixed and Simple Models in Predicting Financial Distress in Firms Listed in Tehran Stock Exchange

The firm financial distress and bankruptcy will result in the waste resources and investment opportunities. In this study, a sample of 210 firms operatingin different studies was surveyed. Of 28 indices under study, 11 indices with the highest impact on the firm financial distress were identified using PartialLeast Squares (PLS). These indices were current assets/current liabilities, current assets/total assets, working capital/ sales, sales/inventory,sales/receivables, net income/liabilities, receivables/ liabilities, net income/ sales, total liabilities/total assets, total liabilities/equities, and currentliabilities/ equities. It was also shown that all independent variables had a significant relationship with the financial distress. In addition, the firmfinancial distress was predicted using the mixed method of Support Vector Machines (SVM) and the simple method (simple SVM). The results of thepaired samples t-test suggested that the mixed SVM is more accurate than the simple SVM in predicting the possibility of the financial distress. It was also noted that the mixed SVM is not only more accurate but also has more generalizability power than the simple SVM

Keywords:

financial distress , partial least squares (PLS) , support vector machines (SVM)