Credit-scoring by enlarged discriminant models

Bankruptcy prediction using support vector machine with optimal choice of kernel function parameters

“…The univariate approach was later improved and extended to multivariate analysis by Altman (1968). During the years that followed, many researchers attempted to increase the success of multiple discriminant analysis (MDA) in predicting business failures (Altman, Marco, & Varetto, 1994;Eisenbeis, 1978;Falbo, 1991;Peel, Peel, & Pope, 1986). Linear probability and multivariate conditional probability models (Logit and Probit) were introduced to the business failure prediction literature in late 1970s.…”

“…So far, linear probability and multivariate conditional probability models, the recursive partitioning algorithm, artificial intelligence, multi-criteria decision making, mathematical programming have been proposed to support the credit decision (Bryant, 1997;Butta, 1994;Cielen & Vanhoof, 1999;Coakley & Brown, 2000;Davis, Edelman, & Gammerman, 1992;Diakoulaki, Mavrotas, & Papayan nakis, 1992;Dimitras, Zanakis, & Zopounidis, 1996; Emel et al, 2003;Falbo, 1991;Frydman, Altman, & Kao, 1985;Jo & Han, 1996;Lee, Sung, & Chang, 1999;Martin, 1997;Reichert, Cho, & Wagner, 1983;Roy, 1991;Tam & Kiang, 1992;Troutt et al, 1996;Zopounidis & Doumpos, 1998). In particular, artificial neural networks (ANNs) are most frequently used in previous literature since the power of prediction is known to be better than the others; however, it has been commonly reported that ANN models require a large amount of training data to estimate the distribution of input pattern, and they have difficulties of generalizing the results because of their overfitting nature.…”

An OWA Operator-Based Cumulative Belief Degrees Approach for Credit Rating

“…1,12 Conventional methods are mainly multivariate statistical methods such as discriminant analysis, regression analysis, logistic regression, and factor analysis. Falbo 13 conducted a discriminant analysis for credit rating. Lee et al 14 took the result of discriminant analysis as data for an input layer for an artificial network and integrated a conventional and a contemporary method while determining credit rating in this manner.…”

“…Conventional methods are mainly multivariate statistical methods such as discriminant analysis, regression analysis, logistic regression, and factor analysis. Falbo conducted a discriminant analysis for credit rating. Lee et al .…”

“…For instance, firm 37's debt ratio (x 3 ) score, which was equal to 1.0646, was transformed the belief degrees by utilizing Equations 13,(14), and (15) as follows. x 1 0.3208 0.6792 1.0000 0.6792 0.0000 0.0000 0.0000 x 2 1.0000 1.0000 0.0000 0.0000 0.0000 0.0000 x 3 0.5646 0.4354 1.0000 1.0000 1.0000 1.0000 0.4354 x 4 0.9100 0.0900 1.0000 0.0900 0.0000 0.0000 0.0000 x 5 1.0000 1.0000 0.0000 0.0000 0.0000 0.0000 x 6 0.1897 0.8103 1.0000 1.0000 1.0000 0.8103 0.0000 x 7 0.3287 0.6713 1.0000 1.0000 1.0000 0.6713 0.0000 x 8 0.6161 0.3839 1.0000 0.3839 0.0000 0.0000 0.0000 x 9 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 x 10 1.0000 1.0000 0.0000 0.0000 0.0000 0.0000 x 11 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 x 12 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 x 13 0.7000 0.3000 1.0000 1.0000 1.0000 0.3000 0.0000 x 14 0.4000 0.4000 0.2000 1.0000 0.6000 0.2000 0.0000 0.0000 x 15 0.2000 0.4000 0.4000 1.0000 1.0000 1.0000 0.8000 0.4000 example, the belief structure of (x 14 ) and (x 15 ) were as follows: Finally, the belief degrees for the firm 37 were calculated as given in Table VIII.…”