Neural nets versus conventional techniques in credit scoring in Egyptian banking

Abstract

Neural nets have become one of the most important tools using in credit scoring. Credit scoring is regarded as a core appraised tool of commercial banks during the last few decades. The purpose of this paper is to investigate the ability of neural nets, such as probabilistic neural nets and multi-layer feed-forward nets, and conventional techniques such as, discriminant analysis, probit analysis and logistic regression, in evaluating credit risk in Egyptian banks applying credit scoring models. The credit scoring task is performed on one bank's personal loans' data-set. The results so far revealed that the neural nets-models gave a better average correct classification rate than the other techniques. A one-way analysis of variance and other tests have been applied, demonstrating that there are some significant differences amongst the means of the correct classification rates, pertaining to different techniques.

An insight into the experimental design for credit risk and corporate bankruptcy prediction systems

“…For example, the ratio of the misclassification cost for type-I error to the misclassification cost for type-II error in the German database was reported to be 5:1 (West 2000), which has further been taken as the ratio between the costs of both errors for other data in a number of papers (Abdou et al 2007(Abdou et al , 2008Abdou 2009b;Lee and Chen 2005). Figure 4 displays the percentages of papers that have employed each of the most typical performance metrics.…”

“…Based on the review carried out, several comments can be outlined: (i) the use of statistical procedures either for determining the optimal method or for comparing the performance of different prediction models appears to be infrequent since more than 68% of papers have not reported any form of hypothesis testing; (ii) the parametric tests have been applied in nearly 18% of papers (especially the t-test with about 15%), but ignoring whether the samples hold the normality and homoscedasticity assumptions or not; (iii) approximately 13% of papers have included a non-parametric test in the experimental protocol, being the McNemar's (5.67%) and Wilcoxon's signed-ranks (3.55%) tests the two most common techniques; (iv) only three papers (Canbas et al 2005;Abdou et al 2008;Abdou 2009a) have studied the statistical difference of variances through Bartlett's, Levene's or Cochran's C tests; and (v) the post hoc tests for comparisons with a control algorithm have seldom been applied, with only seven works using the Tukey's method (Pendharkar 2005), the Nemenyi's test (García et al 2012;Marqués et al 2013;Brown and Mues 2012), the Holm's test (Hu and Chen 2011) or the Bonferroni-Dunn's procedure (Marqués et al 2012a,b).…”

Prediction of financial strength ratings using machine learning and conventional techniques

“…Discriminant analysis (DA) is a classification technique widely used to develop a Z-score model to discriminate between two or more groups of observations (Abdou et al, 2008). DA predicts and classifies problems where the nature of the dependent variable is binary, for example, high versus low risk, high versus low FSRs etc.…”

“…Selected independent variables for the proposed models are reduced to 17 financial and non-financial variables 5 .…”

The Use of Genetic Algorithm, Clustering and Feature Selection Techniques in Construction of Decision Tree Models for Credit Scoring

“…Artificial neural networks (ANNs) [5], naive Bayes, logistic regression(LR), recursive partitioning, ANN and sequential minimal optimization (SMO) [6], neural networks (Multilayer feed-forward networks) [7], ANN with standard feed-forward network [8], credit scoring model based on data envelopment analysis (DEA) [9], back propagation ANN [10], link analysis ranking with support vector machine (SVM) [11], SVM [12], integrating non-linear graph-based dimensionality reduction schemes via SVMs [13], Predictive modelling through clustering launched classification and SVMs [14], optimization of k-nearest neighbor (KNN) by GA [15], Evolutionary-based feature selection approaches [16], comparisons between data mining techniques (KNN, LR, discriminant analysis, naive Bayes, ANN and decision trees) [17], SVM [18], intelligent-agent-based fuzzy group decision making model [19], SVMs with direct search for parameters selection [20], SVM [21], decision support system (DSS) using fuzzy TOPSIS [22], neighbourhood rough set and SVM based classifier [23], Bayesian latent variable model with classification regression tree [24], integrating SVM and sampling method in order to computational time reduction for credit scoring [25], use of preference theory functions in case based reasoning model for credit scoring [26], fuzzy probabilistic rough set model [27], using rough set and scatter search met heuristic in feature selection for credit scoring [28], neural networks for credit scoring models in microfinance industry [29].…”