摘要：Purpose of the article: The purpose of this paper is to analyse the distributional properties of financial data,suitable for building a bankruptcy forecast model,in the sense of normality deviation and the existence of outliers.Methodology/methods: In praxis,financial data in the form of financial ratios is very often not normály distributed.A Shapiro-Wilk’s procedure was used to test normality (Shapiro,Wilk,1965) and a Box-Cox transformation (Box,Cox,1964) for normalizing financial ratios.Scientific aim: We would like to contributed to the previous pieces of research in following ways.Firstly,by analysing a greater range of accounting ratios or indicators (i.e.44),secondly,by focusing on data of a different character (data suitable for building a bankruptcy forecast model),thirdly,by explaining cases in which the parameter l is not possible to estimate,and finally fourthly,identifying a possible cause of transformation failure in achieving normality of financial ratios.Findings: Before the transformation none of the analysed financial ratios met the condition of one-dimensional normality,not even on the 1-% level.After transformation,the condition of one-dimensional normality was met,at the 1-% level,by 34% of the analysed financial ratios.The same condition,but at the 5 or 10-% level,was met by 27% of the analysed financial ratios.The parameter l was not possible to estimate in the case of 18% of financial ratios.Conclusions: The condition of normality for untransformed Czech bankruptcy data seems almost as impossible to fulfil.This conclusion implies the use of non-parametric methods,such as artificial neural networks.However,the comparison of the parametric method’s performance using untransformed or transformed data is the subject of further research.
