Neural network and statistical models for prediction of financial health of an organization

Prediction of sickness of an organization is an important issue in today´s business. It helps an organization to assess the risk involved in getting into a business transaction with a counter party. This is more acute for financial institutions (FIs) due to their nature of business transactions. They need to know the credit risk of an existing loan portfolio to insure it against loss, and of a new loan to determine a fair interest rate. This prediction of sickness of an organization can be made by using soft computing models. In this paper, we have used multi-layer perceptron (MLP) and logistic regression (LR) for predicting the sickness of the public sector enterprises (PSEs) located in India. Simulation results show that MLP performs better than LR in predicting sickness of the PSEs.