Analýza finančního zdraví podniků pomocí matematicko-statistických metod

Abstract

The aim of bachelor thesis is the creation of credit-scoring models for years 2005, 2006 and 2007, respectively in two different versions. First variant is based on twelve ratios and there is available a data sample which includes fourteen banks, while second variant is based on eighteen ratios and models are compiled on the basis of data twelve banks. Linear discriminant analysis is used to formation individual models.

The theoretical part deals with general characteristics of the basic models and financial analysis. There is also included in this part an introduction of individual statistical instruments which are used to formation models. The practical part describes the application of presented theoretical-methodological starting points for the sample banks.

There are six credit-scoring models set up in total. The most reliable model is model for year 2005 compiled on the basis of eighteen ratios. Its success rate is 40 %. Return on average assets has the highest value of coefficients γ, in models which are based on eighteen ratios ratio problem loans to gross loans has the high value of this coefficients as well. These two ratios in compiled models are ratios with the hightest discriminating power.