The Use of Discrete Variable Selections for Credit Evaluation
Omega, Vol. 16, No. 5, pp. 469-480, September 1988
Posted: 18 Apr 2006
A discriminant model for evaluation of loan applications by bank managers should be capable of handling a wide variety of data in a parsimonious manner. Frequently the data will be discrete, and not easily converted to continuous form. Furthermore, missing data may be a problem. This paper examines the use of a discrete discriminant analytic technique in credit screening in a rural setting. The efficiency of the model is evaluated with regard to both rate and costs of misclassification.
Suggested Citation: Suggested Citation