Zoom Training Details:

Day 1:

Duration	Topic
80 min	Current Issues in Credit Risk and LGD Engineering Current issues in credit risk and COVID-19 challenges Introduction into Python Role of LGD and EAD in credit loss modeling Computing observed LGD LGD discount rates Resolution bias
20 min	Break
80 min	Loss Given Default (LGD) Concepts Statistical and machine learning thinking Workflow in machine learning Differences between loss and default modeling Train-test validation split Standardisation and scaling Bias-variance trade-off Cross-validation Hyper-parameters and tuning Linear, fractional logit & beta regressions Feature selection, choce of model and interpretation for LGD PD-LGD models

Day 2:

Duration	Topic
80 min	Exposure Concepts Conversion measures at default and undistressed period Credit lines & flexible repayment schedules Exposures at default Hands-on Excercises in Python: Visualisation and Reporting Communication with non-Python users Presentations in Jupyter Notebook and reveal.js Building user interfaces for recovery management Model deployment in the cloud
20 min	Break
80 min	Machine Learning Validation Validation framework and other continuous variables Backtesting LGDs: discrimination, calibration, stability Adapt accuracy, ROC curves, AUC for LGD and other continuous variables Brier score R^2 Calibration curve and reliability diagram Portfolio dependence

Day 3:

Duration	Topic
80 min	Current expected credit losses (CECL) for US GAAP and IFRS 9 Including LGD and EAD to loan loss provisioning 12-month expected loss vs. lifetime expected loss Macroeconomic forecasts Multi-period forecasts based on the macroeconomy and borrower lifecycles Prediction of lifetime expected losses Significant increase in credit risk (SICR) Cost of debt and capital Competitive and sustainable loan pricing
20 min	Break
80 min	Supervised Machine Learning for LGDs Regression K Nearest Neighbours Decision Trees Bagging and Boosting

Day 4:

Duration	Topic
80 min	Unexpected Credit Losses Downturn LGD and EAD for Basel calibrations and stress-testing Correlation modelling Asymptotic Single Risk Factor Model Numerical integration to generate loss distributions Simulation to generate loss distributions Value-at-Risk and Expected Shortfall
20 min	Break
80 min	Supervised Machine Learning for LGDs Neural Networks Network architecture and depth Deep learning Do it like Google: Introduction into Keras/Tenserflow Support Vector Regressions

---

Back to Zoom Training