EPIA'03 - 11th Portuguese Conference on Artificial Intelligence

	Session: December 6, 11:30-12:15, Room A
Title:	A Data Mining Approach to Credit Risk Evaluation and Behaviour Scoring
	Sara C. Madeira, Arlindo L. Oliveira, Catarina S. Conceição
Abstract:	Nowadays, behaviour scoring is used in several companies to score the customers according to credit risk by analyzing historical data about their past behaviour. In this paper we describe a data mining approach to credit risk evaluation in a Portuguese telecommunication company. The predictive models obtained, which were derived using recent data from the data warehouse of the company and the data mining tool SAS Enterprise Miner, infer the credit risk of the customers three months in the future given six months of historical data. Standard machine learning techniques, like statistical regression, neural networks and decision trees, were used. The credibility and precision of this widely used techniques, together with the possibility to predict the default of the customers three months in advance, are the keys to the success of this approach to credit risk evaluation.

Back to schedule.