Predicting Bankruptcy With Precision: Insights From Hybrid Machine Learning Models On Unbalanced Polish Financial Data

• Author(s):

  Aishwarya M | Anu Shree R M | Dhanyashree T N | Meghana Shree S | Dr.Vishwesh J

• Keywords:

  Bankruptcy Forecasting, Deep Learning, Ensemble Methods, Hybrid Machine Learning, Imbalanced Dataset

• Abstract:

  Bankruptcy Prediction Is A Critical Area In Financial Risk Assessment, Supporting Timely Decisions For Investors, Regulators, And Institutions. This Study Presents A Comparative Analysis Of Multiple Machine Learning Models, Including Traditional Algorithms (Decision Tree, Naive Bayes), Deep Learning Methods (CNN, LSTM), And Hybrid Approaches (XGBoost + ANN, Decision Tree + Gaussian), Applied To An Imbalanced Financial Dataset From Polish Companies. The Dataset Poses Real-world Challenges Such As Class Imbalance And Feature Noise, Which Are Addressed Through Data Preprocessing, Feature Selection, And Resampling Techniques. The Proposed Hybrid Models Integrate The Strengths Of Ensemble Learning And Neural Networks, Improving Classification Performance On Minority (bankrupt) Classes. Evaluation Using Metrics Like Precision, Recall, And F1-score Demonstrates That Hybrid And Deep Learning Models Outperform Traditional Classifiers, With The XGBoost–ANN Model Achieving The Best Overall Results. Feature Importance Analysis Further Reveals The Most Influential Financial Indicators Contributing To Bankruptcy Prediction. This Work Offers A Robust, Adaptable Framework For Handling Imbalanced Datasets In Financial Domains, Contributing Practical Insights For Early Risk Detection And Decision-making.

---

Other Details

• Paper id:

  IJSARTV11I5103465

• Published in:

  Volume: 11 Issue: 5 May 2025

• Publication Date:

  2025-05-05

Download Article