Content area

Abstract

This paper aims to identify some financial characteristics that differentiate healthy companies from failing ones and evaluate the performance of five classification methods for predicting the business failure of Moroccan firms: LDA, LR, SVM, ANN, and K-NN. Using 51 healthy and 40 failing companies in 2019, it was found that failing firms, compared with healthy companies, exhibited lower profitability, limited internal generation of resources, high dependence on supplier credit, and lengthy recovery periods. Results from K-fold cross-validation show that LDA can detect failures but returns many false positives; ANNs balance accuracy with sensitivity well. SVM reduces false positives but could miss some failures, while K-NN and LR are unpredictable with complex data. Bootstrapping necessitates model generalisation, where ANNs are the most acceptable model for classification-based prediction. The study offers valuable insights for stakeholders to enhance financial risk assessments despite limitations such as sample size and static analysis.

Details

10000008
Business indexing term
Location
Title
Evaluating Financial Prediction Models for Business Failure in Moroccan Industrial Firms: Analysis and Strategic Implications
Volume
20
Issue
2
Pages
171–196
Number of pages
27
Publication year
2025
Publication date
Jun 2025
Publisher
Adonis & Abbey Publishers Ltd
Place of publication
London
Country of publication
United Kingdom
Publication subject
ISSN
17504554
e-ISSN
17504562
Source type
Scholarly Journal
Language of publication
English
Document type
Journal Article
Publication history
 
 
Online publication date
2025-06-30
Publication history
 
 
   First posting date
30 Jun 2025
ProQuest document ID
3236094427
Document URL
https://www.proquest.com/scholarly-journals/evaluating-financial-prediction-models-business/docview/3236094427/se-2?accountid=208611
Copyright
Copyright Adonis & Abbey Publishers Ltd 2025
Last updated
2025-11-07
Database
ProQuest One Academic