چکیده:
In this paper we construct a modeling for detection of banks which are experiencing serious problems. Sample and variable set of the study contains 30 banks of Iran during 2006-2014 and their financial ratios. Well known multivariate statistical technique (principal component analysis) was used to explore the basic financial characteristics of the banks, and discriminant Logit and Probit models were estimated based on these characteristics. Results suggest that the model can be used as an analytical decision support tool in both on-site and off-site bank monitoring system to detect the banks which are experiencing serious problems.
خلاصه ماشینی:
"Well known multivariate statistical technique (principal component analysis) was used to explore the basic financial characteristics of the banks, and discriminant Logit and Probit models were estimated based on these characteristics.
2. Literature Review Previous bank studies which employed multivariate statistical analysis include discriminant model (Sinkey, 1975), Logit models [Rose and Kolari, 1985; Pantolone and Platt, (1987)], and Probit model (Cole and Gunther, 1998).
Some other new studies tend to combine the non-parametric approaches with the discriminant or Logit analysis for bank failure prediction; Tam and Kiang (1992) introduced neural network approach to perform discriminant analysis as a promising method of evaluating bank conditions.
Well known multivariate statistical technique (Principal Component Analysis), was used to explore the basic financial characteristics of banks, and discriminant, Logit and Probit models were estimated based on these characteristics to construct IEWS.
(2009) aimed to apply various neural network techniques, support vector machines and multivariate statistical methods to the bank failure prediction problem in a Turkish case to present a comprehensive computational comparison of the classification performances of the techniques tested.
/ / In the following sections, Principal Component Analysis (PCA) was applied to 23 early warning ratios and the important factors for explaining changes in financial conditions of bank were determined.
6F5 (7) lb b b b b b Table 7: Test Statistics for the Estimated Logit and Probit Models رجوع شود به تصویر صفحه Source: Research findings In the Probit method the probability ( P ) of a bank going to failure is b given by cumulative standard normal distribution function: PP Z Pb 1 z e 2 dz ."