A review of bankruptcy prediction models: The machine learning perspective

For financial institutions, it is very important to have the ability to predict or forecast business failure. Incorrect decision making is likely to cause financial distress or crises. Bankruptcy prediction is a major research problem in the accounting and finance domain. In other words, it would be very useful if financial institutions have a prediction model which is able to predict whether the loan customers would be bankrupt or not. Using machine learning techniques, such as neural networks, decision trees, etc. to develop bankruptcy prediction models has been extensively studied since the 1990s. This chapter reviews 27 related journal papers in the period between 2000 and 2007 focusing on developing hybrid and ensemble classifiers. Related studies are compared by their classifier design, datasets used, and other experimental setups. Current achievements and limitations in developing bankruptcy prediction models by machine learning are present and discussed. A number of future research directions are also provided.