Credit model

Credit Risk = PD * EAD * LGD (Credit Risk = Probability of default * Exposure at default * Loss given default)

Bankruptcy prediction

feature reduction: recursive feature selection with cross validation (RFECV)
classifier models applied: random forest, gradientboost, xgboost, lightGBM
hyperparameter optimization: RandomSearchCV, GridSearch
metrics for model evaluation: confusion matrix, ROC-AUC, accuracy, precision, recall, f1-score, support
metrics for feature importance: feature importance, Shapley values, Morris sensitivity

Name		Name	Last commit message	Last commit date
Latest commit History 83 Commits
.github/workflows		.github/workflows
.ipynb_checkpoints		.ipynb_checkpoints
img		img
ss		ss
.DS_Store		.DS_Store
CreditRisk.pdf		CreditRisk.pdf
IV calculation.ipynb		IV calculation.ipynb
K-S calculation.ipynb		K-S calculation.ipynb
PSI calculation.ipynb		PSI calculation.ipynb
Predicting bankruptcies Pipeline.ipynb		Predicting bankruptcies Pipeline.ipynb
Predicting bankruptcies.ipynb		Predicting bankruptcies.ipynb
README.md		README.md
auroc_test.csv		auroc_test.csv
test_iv.csv		test_iv.csv
test_ks.csv		test_ks.csv
test_psi.csv		test_psi.csv