Decision Lists Applied to the Bankruptcy Prediction in Non-Life Insurance Sector.

Abstract

Prediction of insurance companies insolvency has arisen as an important problem in the field of financial research, in order to protect both society and customers and minimize the costs associated with this problem.
The development and application of new criteria for a better harnessing of the financial-accounting information furnished by insurance companies is a key issue to handle this problem. In line with this issue, this paper aims to examine the applicability of a technique coming from the Artificial Intelligence to the prediction of insolvency in the insurance sector, a learning algorithm for decision lists called PART. We also compare this method with a classical statistical approach, Logistic Regression. PART presents the advantage of being easy to implement as well as providing results of simple interpretation whilst avoiding some of the drawbacks of the conventional statistical techniques.