πŸ“‰

Bankruptcy Prediction

Project Description

In this project, I worked with financial data from the Emerging Markets Information Service. I looked at financial indicators from Poland and Taiwan, and built supervised machine learning models (random forests & gradient boosting) that predict bankruptcy.

Overview

  • Addressed imbalanced data using resampling techniques
  • Evaluated models using classification metrics using precision and recall
  • Improved bankruptcy models using random forests and gradient boosting
  • Evaluated model generalizability using cross validation
  • Tuned model hyperparameters using grid search
  • Built prediction functions that use saved models
  • Created Python modules to store prediction functions

Language & Tools

  • Python (Pandas, Matplotlib, Seaborn, Scikit-learn)

Featured Notebooks