Application of credit scoring models in electricity companies

Electricity companies face great credit risks because of the consume-and-pay method of electricity payment. Credit scoring is a very important method in recognizing credit risk, so this study investigates the classification models to identify the credit risk inherent in the payment method used by electricity companies. Three different classification methods, i.e. decision tree, neural networks and logistic regression, are examined for their suitability in credit scoring. As the results reveal, logistic regression outperforms the other alternatives. This paper presents a useful framework to choose the best model to recognize the credit risk for electricity companies.