Forecasting company financial distress: C4.5 and adaboost adoption

Abstract

Financial pressure is one of the factors that determine the survival of a business. In order to minimize the exhausted risk, economic-financial analytics and forecasting have been taken into account. Therefore, this study aims to cover the Altman model to monitor and assess the financial situation based on the financial report’s balance sheet and income statement to predict the financial distress status into Health, Undefined, and Distress condition. Here, the integration of the C4.5 algorithm and Adaboost carried out five Altman’s worth attributes for optimally undermining the financial distress index, which includes working capital to total assets (X1), retained earnings to total assets (X2), earnings before interest, and taxes to total assets (X3), market value of equity to book value of total liabilities (X4) and sales to total assets (X5). Furthermore, the Knowledge of Data Discovery (KDD) executed 755 data records of financial reports from the Indonesia Stock Exchange during the Year 2016-2019 to analyze its accuracy and error rate using this combining approach. The Confusion Matrix showed that algorithms C4.5 and AdaBoost forecast were 13.52% and 62.17% more precise than the original C4.5 and Altman’s model, respectively, in ratio training tested data 90%:10%. This study, therefore, revealed the substantial contribution of C4.5 and Adaboost to company financial distress forecasting.

Keywords

• adaboost
• c4.5 algorithm
• financial distress
• datamining
• prediction