Forecasting with neural networks. An application using bankruptcy data

Desmond Fletcher, Ernest Goss

Research output: Contribution to journalArticle

251 Citations (Scopus)

Abstract

In the business environment, Least-Squares estimation has long been the principle statistical method for forecasting a variable from available data with the logit regression model emerging as the principle methodology where the dependent variable is binary. Due to rapid hardware and software innovations, neural networks can now improve over the usual logit prediction model and provide a robust and less computationally demanding alternative to nonlinear regression methods. In this research, a back-propagation neural network methodology has been applied to a sample of bankrupt and non-bankrupt firms. Results indicate that this technique more accurately predicts bankruptcy than the logit model. The methodology represents a new paradigm in the investigation of causal relationships in data and offers promising results.

Original languageEnglish
Pages (from-to)159-167
Number of pages9
JournalInformation and Management
Volume24
Issue number3
DOIs
StatePublished - 1993

Fingerprint

Neural networks
Backpropagation
Statistical methods
Innovation
Hardware
Bankruptcy
Methodology
Industry
Paradigm
Nonlinear regression
Prediction model
Software
Logit
Least squares
Regression model
Logit model
Regression method
Logit regression
Back-propagation neural network
Business environment

All Science Journal Classification (ASJC) codes

  • Computer Science (miscellaneous)
  • Management Information Systems
  • Information Systems
  • Information Systems and Management

Cite this

Forecasting with neural networks. An application using bankruptcy data. / Fletcher, Desmond; Goss, Ernest.

In: Information and Management, Vol. 24, No. 3, 1993, p. 159-167.

Research output: Contribution to journalArticle

@article{424cfab147fc46e8950c51c711288283,
title = "Forecasting with neural networks. An application using bankruptcy data",
abstract = "In the business environment, Least-Squares estimation has long been the principle statistical method for forecasting a variable from available data with the logit regression model emerging as the principle methodology where the dependent variable is binary. Due to rapid hardware and software innovations, neural networks can now improve over the usual logit prediction model and provide a robust and less computationally demanding alternative to nonlinear regression methods. In this research, a back-propagation neural network methodology has been applied to a sample of bankrupt and non-bankrupt firms. Results indicate that this technique more accurately predicts bankruptcy than the logit model. The methodology represents a new paradigm in the investigation of causal relationships in data and offers promising results.",
author = "Desmond Fletcher and Ernest Goss",
year = "1993",
doi = "10.1016/0378-7206(93)90064-Z",
language = "English",
volume = "24",
pages = "159--167",
journal = "Information and Management",
issn = "0378-7206",
publisher = "Elsevier",
number = "3",

}

TY - JOUR

T1 - Forecasting with neural networks. An application using bankruptcy data

AU - Fletcher, Desmond

AU - Goss, Ernest

PY - 1993

Y1 - 1993

N2 - In the business environment, Least-Squares estimation has long been the principle statistical method for forecasting a variable from available data with the logit regression model emerging as the principle methodology where the dependent variable is binary. Due to rapid hardware and software innovations, neural networks can now improve over the usual logit prediction model and provide a robust and less computationally demanding alternative to nonlinear regression methods. In this research, a back-propagation neural network methodology has been applied to a sample of bankrupt and non-bankrupt firms. Results indicate that this technique more accurately predicts bankruptcy than the logit model. The methodology represents a new paradigm in the investigation of causal relationships in data and offers promising results.

AB - In the business environment, Least-Squares estimation has long been the principle statistical method for forecasting a variable from available data with the logit regression model emerging as the principle methodology where the dependent variable is binary. Due to rapid hardware and software innovations, neural networks can now improve over the usual logit prediction model and provide a robust and less computationally demanding alternative to nonlinear regression methods. In this research, a back-propagation neural network methodology has been applied to a sample of bankrupt and non-bankrupt firms. Results indicate that this technique more accurately predicts bankruptcy than the logit model. The methodology represents a new paradigm in the investigation of causal relationships in data and offers promising results.

UR - http://www.scopus.com/inward/record.url?scp=0000152448&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0000152448&partnerID=8YFLogxK

U2 - 10.1016/0378-7206(93)90064-Z

DO - 10.1016/0378-7206(93)90064-Z

M3 - Article

AN - SCOPUS:0000152448

VL - 24

SP - 159

EP - 167

JO - Information and Management

JF - Information and Management

SN - 0378-7206

IS - 3

ER -