Publication detail

A Parametric or Nonparametric Approach for Creating a new Bankruptcy Prediction Model: The Evidence from the Czech Republic

KARAS, M. REŽŇÁKOVÁ, M.

Original Title

A Parametric or Nonparametric Approach for Creating a new Bankruptcy Prediction Model: The Evidence from the Czech Republic

Type

journal article in Scopus

Language

English

Original Abstract

For many years now the development of models capable of predicting company bankruptcy has aimed at increasing their accuracy. Among the decisive factors determining the accuracy of the bankruptcy model have been the choice of variable models and applied classification algorithms. The prevailing opinion in literature is that the accuracy of bankruptcy models cannot be appreciably improved by the choice of classification algorithm. A reflection of this assertion is the frequent usage of parametric methods. In particular this involves the method of linear discrimination analysis. This method formed the basis of the first bankruptcy model and continues to be the most frequently applied classification algorithm. However, it requires the fulfilment of assumptions which financial data does not provide and therefore limits the improvement of the models predictive capabilities. This led the authors to the idea of testing the possibility of improving the bankruptcy models predictive capabilities by using non-traditional approaches. Using the example of companies from the Czech Republic it was discovered that a nonparametric method, when used for the selection of model variables as well as the actual classification, can yield significantly better results than the traditional parametric approach.

Keywords

bankruptcy prediction models, boosted trees, stepwise discriminant analysis.

Authors

KARAS, M.; REŽŇÁKOVÁ, M.

RIV year

2014

Released

12. 6. 2014

ISBN

1998-0140

Periodical

INTERNATIONAL JOURNAL of MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

Year of study

8

Number

1

State

United States of America

Pages from

214

Pages to

223

Pages count

10

BibTex

@article{BUT107879,
  author="Michal {Karas} and Mária {Režňáková}",
  title="A Parametric or Nonparametric Approach for Creating a new Bankruptcy Prediction Model: The Evidence from the Czech Republic",
  journal="INTERNATIONAL JOURNAL of MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES",
  year="2014",
  volume="8",
  number="1",
  pages="214--223",
  issn="1998-0140"
}