Predicting corporation bankruptcy using cash flow statement: applying artificial neural network

Financial crisis of big companies during this decade led to introduce bankruptcy prediction models. The main objective of this study was to assess the information content of the cash flow of the companies listed in Tehran Stock Exchange bankruptcy detection using artificial neural networks. The population of this research firms listed in Tehran Stock Exchange for the period from 2005 to 2013 is the year. For this purpose, 84 companies including 42 companies’ bankrupt and 42 healthy firms were selected. This three-layer perceptron neural network is trained using back-propagation algorithm. According to the results, the neural network model to current liabilities ratio of operating cash flow, operating cash flow to interest coverage ratio, cash return on assets ratio, the ratio of earnings quality expected instantaneous power in Iran is capable to bankruptcy. The findings also show that the model accurately predicted 99 percent of the total for the year of bankruptcy, insolvency procedures in two or three years before bankruptcy 91 and 85 and 70%, respectively, accurately predicts.