🏦 Bankruptcy Prediction Using Machine Learning

This project aims to predict whether a company will go bankrupt based on various financial risk factors. Multiple classification models were trained, compared, and evaluated using precision, recall, and F1-score, with a focus on minimizing false negatives.

---

🔍 Project Overview

• Loaded and explored bankruptcy dataset
• Handled data imbalance between bankrupt and non-bankrupt classes
• Preprocessed data using StandardScaler and LabelEncoder
• Compared multiple classification models:
  • Logistic Regression
  • Gaussian Naive Bayes
  • K-Nearest Neighbors (KNN)
  • Decision Tree
  • Random Forest
  • Gradient Boosting Classifier
• Tuned hyperparameters using cross-validation and GridSearchCV
• Evaluated models using:
  • Confusion Matrix
  • Accuracy, Recall, Precision, F1-score
• Selected Gradient Boosting as the best model with:
  • Accuracy: 99%
  • Recall: 98.7%
  • F1-score: 99%

---

📊 Technologies Used

• Python (Jupyter Notebook)
• Pandas, NumPy
• Scikit-learn (classification models, preprocessing, metrics)
• Seaborn, Matplotlib (for visualizations)

---

📈 Key Results

• Achieved high performance across metrics with Gradient Boosting
• Correctly handled imbalanced classes with strategy and metric focus
• Built a reliable bankruptcy prediction tool for financial risk modeling

---

📁 Files Included

File	Description
BANKRUPTCY_PREDICTION.ipynb	Full notebook with all models, evaluation, and results
bankruptcy.csv (optional)	Dataset (if applicable)
README.md	This documentation file