Application of Artificial Neural Network Models in Predicting Economic Bankruptcy of the Registered Corporations in Stock Market

Authors

1 Faculty member of Tehran University

2 M. A. in Economics

Abstract

"Artificial Neural Network" is known as one of the most advanced "Bankruptcy Prediction Models". Based on this research paper، the main structure of three - and four - layer Perceptron models for bankruptcy prediction had led to same models; however، three-layer network has benefited from more predictive power than four - layer network. In comparison with other models، applying neural network models can improve the potentials of financial managments to stand against economic fluctuations and bankruptcy. Through this paper، the economic bankruptcy of registered corporations in Tehran Stock Market during ٢٠٠٦ – ٢٠٠٧ would be predicted and bankruptcy trend of above- mentioned corporations between ١٩٩١ - ٢٠٠٧ have been presented analytically. As a result of defining economic policies، the trend of economic bankruptcy would be remarkably increased through ٢٠٠٦. A year later، ٢٠٠٧، due to the compatibility of corporations with present situations، the above – mentioned trend would be adjusted.