Adjustment of the Iterative Reclassification Method for Including the Rejected Applications into the Credit Scoring

Олександр Миколайович Солошенко


The objective of the research is the adjustment of the Iterative Reclassification Method for including the rejected applications into the credit scoring. The methodology of implementation uses partially classified data and the logistic regression generalization. The method of the adjustment the Weight Of Evidence and the Information Value indicators using the rejected loan applications is proposed at the first stage. The process of including the rejected applications into predictive power analysis of characteristics has been demonstrated, providing alternative and adjusted discretization process for the continuous variables. The method of including the adjusted Weight Of Evidence and the partially classified rejected applications into logistic regression procedure is proposed at the second stage. The general method of the Iterative Probability Recalculation is proposed at the final stage, using the adjusted logistic regression approach. The research results are the significant improvement of the Iterative Reclassification Method and the generalization of the logistic regression application. As a conclusion, the main advantages of the method are given in comparison with the classical Iterative Reclassification Method, for instance, the probabilistic duality of the rejected applications is mentioned here.


Credit scoring; Reject inference; Logistic regression; Weight Of Evidence, Information Value, partially classified data, data mining, binary classification; Iterative Reclassification Method


Энциклопедия финансового риск-менеджмента: Учеб. пособие / Под ред. А.А. Лобанова и А.В. Чугунова. – М.: Альпина Паблишер, 2003. – 786 с.

Lyn C. Thomas et al., Credit Scoring and its Applications: SIAM monographs on mathematical modeling and computation. Philadel­phia: University City Science Cen­ter, SIAM, 2002, 248 p.

Naeem Siddiqi. Credit risk scorecards: developing and implementing intelligent credit scoring. Hoboken: John Wiley & Sons, Inc., 2006, 196 p.

A.J. Feelders “Credit scoring and reject inference with mixture models”, Int. Journal of Intel. Sys. in Ac­counting, Finance and Management, no. 8, pp. 271–279, 1999.

Jie-Men Mok, Reject Inference in Credit Scoring. Amsterdam: BMI paper, 2009, 38 p.

Руководство по кредитному скорингу: Учеб. пособие / Под ред. Элизабет Мэйз; пер. с англ. И.М. Тикота; науч. ред. Д.И. Вороненко. – Минск: Гревцов Паблишер, 2008. – 464 с.

Paul D. Allison, Logistic Regression Using the SAS® System: Theory and Application. Cary, NC: SAS Institute Inc., North Carolina, 1999, 306 p.

Бідюк П.І., Кузнєцова Н.В., Терентьєв О.М. Система підтримки прийняття рішень для аналізу фінансових даних // Наук. вісті НТУУ “КПІ”. – 2011. – № 1. – С. 48–61.

GOST Style Citations



Copyright (c) 2017 NTUU KPI