PulseTrendingJournal ClubResearchersJournalsExplore
Instagram
HomeTrendingJournal ClubExplore
Synapse
⌘+K
Synapse
February 2, 2026Journal of Accounting & Organizational ChangeOpen Access

Corporate financial distress prediction: a machine learning approach in the era of big data

View Full Paper
Ask AI
Bookmark
Share

Authors

GGGianluca GabrielliAMAndrea MelioliFBFlavio Bertini

Discussion

Loading...

Member takes

Overview

Demonstrates enhanced bankruptcy prediction in Italian firms through modern machine learning approaches, suggesting better forecasting methods.

Key Points

  • The aim is to evaluate the effectiveness of machine learning classifiers in predicting corporate bankruptcy using extensive financial data.
  • Analyzed 1,826,157 firm-year observations from 1980 to 2019.
  • Compared various machine learning models, including random forest and gradient boosting.
  • Addressed class imbalance through resampling techniques like SMOTE.
  • Validated models on held-out samples, regional subsets, and out-of-time tests.
  • Ensemble methods outperformed other classifiers with AUCs near 0.99.
  • F1-scores reached up to 0.98 with raw accounting inputs.
  • Resampling techniques enhanced model robustness.
  • Capital-structure metrics emerged as critical early warning indicators.

Cite This Study

Gabrielli et al. (2025) studied this question.

synapsesocial.com/papers/6980fe57c1c9540dea810641https://doi.org/10.1108/jaoc-05-2025-0166
View Full Paper
Ask AI
Bookmark
Share