Predicting Iranian companies bankruptcy by the use of artificial neural network and art map fuzzy neural network

Investors, creditors and other users of financial information always cope with the risk of losing actual or potential financial resources in their decision makings, thus everything which could eliminate the above mentioned risk or even reduce it has high importance for them and it has a special place. One of the mechanisms that could help the investors and others in reducing the investment risk is paying attention to the accounting information and patterns of predicting bankruptcy. Therefore the current research studies the ability of two models of predicting bankruptcy in Iran Stock Market during t, t-1 and t-2 years. The t year is considered as the bankruptcy year for the bankrupt companies and it is considered as the year of being placed in the sample for the non-bankrupt companies. The current research sample includes 110 accepted companies in Stock Exchange which had the determined conditions. The sampling interval is from 2003 to 2006 and the required data extraction interval includes years 2001 to 2006. The research result shows that each of the two models have the skill of predicting the bankruptcy of companies that are members of Stock Exchange but the performance of ART MAP fuzzy neural network is better.