Recombining Forecasts Used in Personal Credit Scoring

Using the idea of combining forecasts, this paper presents a new approach by combining multi-linear regression and logistic with three different NNs. Then, recombine them with perceptron and apply it in personal credit scoring of one commercial bank. The results indicate that the combining methods are more accurate than either of the individual technology, and recombining is a reasonable way because it has greater precision than either the combining methods or pre-combining models, especially in avoiding recognizing the bad applications as good ones