Financial Distress Classification Using Feedforward Neural Network Based on Altman and Ohlson Financial Ratios

Abstract

The ever-changing economy requires companies to anticipate future conditions in order to avoid financial distress, a continuous decline in financial conditions. The research focused on comparing Altman and Ohlson’s financial ratio in classifying financial distress on Property and Real Estate companies using the Feedforward Neural Network. The data used is the financial report data of 19 Property and Real Estate companies listed on the Indonesian Stock Exchange in 2016-2022, with the initial status of financial conditions based on earnings per share. (EPS). The study also used the Synthetic Minority Oversampling Technique (SMOTE) method to address class imbalances. The best financial ratio is selected based on accuracy values and Area Under Curve (AUC). Altman’s financial ratio with the FFNN model architecture (5-2-1) with a balance of 60:40 yields an accuracy of 84.62% and an AUC of 0.8325. The Ohlson Financial ratio with the 60:40 data balancing process and the FFNN model architecture (9-4-1) yields an accuracy of 93.27% and an AUC of 0.9045. Thus, in predicting financial distress in companies in the Property and Real Estate sector, Ohlson’s financial ratio with the predictor variables Corporate Size (SIZE), Total Liabilities to Total Assets (TLTA), Working Capital to Total Acts (WCTA), Current Liability to Current Asset (CLCA), OENEG, Net Income to total assets (NITA), Cash Flows Operating to Total Responsibilities (CFOTL), Net Revenue (INTWO), and Net Incoming Change (CHIN) yielded the best results. This best ratio can be used as a consideration in using alternative financial ratio to classify financial distress.