Predicting bankruptcy using neural networks and other classification methods: The influence of variable selection techniques on model accuracy

Jardin, Philippe du

doi:10.1016/j.neucom.2009.11.034

Cited by 111 publications

(51 citation statements)

References 67 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Bármelyikről is legyen szó az említettek közül, abban a tekintetben mindenkép-pen egységesnek tekinthető a szakirodalom, hogy a kutatási terület célja a modellek előrejelző képességének növelése (Du Jardin [2010]). Tanulmányunkban magunk is e célt tűztük magunk elé.…”

Section: Munkáját)unclassified

A pénzügyi mutatók időbeli tendenciájának figyelembevétele logisztikus regresszióra épülő csődelőrejelző modellekben

Nyitrai¹,

Virág²

2017

Statisztikai Szemle

View full text Add to dashboard Cite

A vállalatok jövőbeli fizetőképességének előrejel-zésében általános gyakorlat a számviteli adatok alapján kalkulált hányados típusú pénzügyi mutatók használata magyarázóváltozóként. A hazai és a nemzetközi szakirodalom általános sajátosságának mondható az is, hogy e mutatókat csak az utolsóként megfigyelt üzleti év adatai alapján kalkulálják. Az e változókra épített modellek azonban statikus jellegűek, melyek nem veszik figyelembe a vállalati gazdálkodás folyamatjellegét. E hiány pótlására korábban Nyitrai [2014a] tett kísérletet a statikus pénzügyi mutatószámok idősorai-ból képzett dinamikus pénzügyi mutatók segítségével. Az idézett kutatás azonban csak a döntési fák vonatkozásában igazolta a dinamikus mutatók hatékonyságát a csődelőrejelzésben. A szerzők tanulmányukban bemutatják a dinamikus változók teljesítményét a gyakorlati hitelkockázati modellezésben általánosan elterjedt logisztikus regresszió keretei között, továbbá kísérletet tesznek a dinamikus pénzügyi mutatók koncepciójának továbbfejlesztésére előrejelző képességük növelése érdekében. TÁRGYSZÓ:Csődelőrejelzés. Dinamikus modell. Logisztikus regresszió.

show abstract

Section: Munkáját)unclassified

A pénzügyi mutatók időbeli tendenciájának figyelembevétele logisztikus regresszióra épülő csődelőrejelző modellekben

Nyitrai¹,

Virág²

2017

Statisztikai Szemle

View full text Add to dashboard Cite

show abstract

“…In order to ensure practical learning of students and understanding the impact of financial ratio on business efficiency, some models assessing the prediction power of financial ratios will be included on the online platform. The models proposed use various prediction techniques: robust logistic regression [14], discriminant analysis [15] [16], data development analysis and multi-layer perception [17], neural networks [18], hazard models [19] [21], logistic models [19][20] [21] and Bayesian models [19] [22]. More models will be grounded based on extensive literature review and the study developed by Bellovary, Giacomino and Akers [22].…”

Section: Concept Of the E-learning Module On Business Failurementioning

confidence: 99%

Concept: Using E-Learning Platforms in Economical and Entrepreneurial Development

Silvestru¹,

Burcezan²,

Bere³

et al. 2015

View full text Add to dashboard Cite

“…In most general terms artificial neural networks is a nonlinear approach, which combines input data through different layers into a single output. As demonstrated in [20], numerous different approaches have been used for artificial neural network creation, the multilayer perceptron trained using a back-propagation method being most widely applied option. Still, genetic algorithm as method should lead to better results, but is more time consuming [25].…”

Section: Summary Of Relevant Literaturementioning

confidence: 99%

“…As different novel methods have positive and negative sides, none of them can directly be considered superior of others. In recent years, a myriad of bankruptcy prediction studies has been conducted based on neural networks (see [20]), which makes it most Predicting Bankruptcy of Manufacturing Firms…”

Section: Summary Of Relevant Literaturementioning

confidence: 99%

Predicting Bankruptcy of Manufacturing Firms

Grünberg¹,

Lukason²

2014

IJTEF

View full text Add to dashboard Cite

Abstract-This paper aims to create bankruptcy prediction models using logistic regression and neural networks based on the data of Estonian manufacturing firms. The models are composed and tested on the whole population data of bankrupt firms and their vital counterparts for years [2005][2006][2007][2008]. Composed models are also tested on the data of firms from economic recession years of 2009-2010. The results indicate that models based on different methods have similar predictive abilities, yet two and three years before bankruptcy they are not as good as for one year before bankruptcy. Also, the models do not perform as well when using data from economic recession years.Index Terms-Bankruptcy prediction, manufacturing firms. I. INTRODUCTIONSince 1960ies the failure prediction domain in literature has flourished. The main idea of failure prediction studies is to establish decision rules based on a set of variables (commonly financial ratios), which would facilitate to discriminate vital and failing firms. During past decades many literature reviews (e.g. [1]- [6]) have appeared, which list a myriad of different prediction studies and the amount of relevant research seems to be quickly increasing. The innovativeness of emerging studies mostly lies in novel statistical techniques, leaving other possibilities to contribute to literature in the background.Still, for increasing the validity of a prediction model, more attention should be directed to the data applied. Therefore, this paper does not aim to make a contribution by elaborating a new statistical technique for failure prediction, but instead addresses to major limitations concerning the data used in prediction model composition. The objective of the paper is to compose bankruptcy prediction models by using one classical and one modern statistical technique, at the same time addressing a set of known data limitations. The novelty of the paper rises from the objective, namely an elaborate dataset will allow addressing multitude of limitations that are rarely viewed together in available empirical studies. As a classical technique, logistic regression, and as a modern technique, neural networks will be applied in empirical analysis. Because of that, literature review is mostly focused on given two methods and less attention is directed to others.The paper is structured as follows. The introductory part is followed by a short review of literature, which specifically focuses on two methods chosen to model failure in current study and in addition addresses some of the general features of bankruptcy prediction models. The following empirical Manuscript received July 20, 2013; revised October 15, 2013. Oliver Lukason and Martin Grünberg are with the Tartu University, Faculty of Economics and Business Administration, Narva road 4, Tartu 51009, Estonia (e-mail: oliver.lukason@ut.ee).analysis is broken into two parts, of which the first addresses data and methodology, whereas the second outlines results of statistical analysis with relevant comments and discussion...

show abstract

Predicting bankruptcy using neural networks and other classification methods: The influence of variable selection techniques on model accuracy

Cited by 111 publications

References 67 publications

A pénzügyi mutatók időbeli tendenciájának figyelembevétele logisztikus regresszióra épülő csődelőrejelző modellekben

A pénzügyi mutatók időbeli tendenciájának figyelembevétele logisztikus regresszióra épülő csődelőrejelző modellekben

Concept: Using E-Learning Platforms in Economical and Entrepreneurial Development

Predicting Bankruptcy of Manufacturing Firms

Contact Info

Product

Resources

About