Bankruptcy Prediction Using Artificial Neural Network
Bankruptcy Prediction Using Artificial Neural Network
Sanaz Sarafraz1 Farid Sefati2 Ali Reza Ghiasvand3 Masoumeh Tabatabaie4
1) Department of financial management, Boroujerd Branch ,Islamic Azad University, Boroujerd, Iran
2) 2,3Department of Accounting, Boroujerd Branch, Islamic Azad University, Boroujerd, Iran
3) 2,3Department of Accounting, Boroujerd Branch, Islamic Azad University, Boroujerd, Iran
4) Young Researchers and Elite Club,Boroujerd Branch , Islamic Azad University, Boroujerd, Iran
IOI Code
:
XYFC-BBBZG
Publication :
International Conference on new Idea In Management Economics & Accounting(imeaconf.com)
Abstract :
Companies’ bankruptcy is one of the important economical phenomena. Health and success of the companies are the major concerns of policy makers, industrialists, investors and managers. In this study, the economic bankruptcy of companies in the year t isis predicted by the multilayer perceptron model of artificial neural network. The financial ratios of Altman z index have been used as independent variables, and healthy and bankrupt companies are dependent variables. The population includes information of companies’ financial statements in the Tehran Stock Exchange during the year 2014. The bankrupt group is selected based on article 141 of the Commercial Code and net loss, and non-bankrupt group is chosen based on profitability criteria. The result of spss 20 software for multilayer perceptron includes 6 input layer, 4 hidden layer, and 2 output layer. And model accuracy in training is 88.75% and is equal to 86.97% for the test.
Keywords : bankruptcy prediction multilayer perceptron article 141 of the commercial code