Zoom Training Details:

Day 1:

Duration	Topic
80 min	Data Pre-processing and COVID-19 Introduction to Python, data and code Outlier detection Standardisation & categorisation Default definitions and drivers (borrower equity and liquidity) Non-monotone relations between defaults and risk factors Weight-of-evidence using the WOE() function, categorisation, splines Time-vintage-age analysis Dataprep() function COVID-19 challenges
20 min	Break
80 min	Probabilities of Default (PD)– Expected PDs Discrete time models (Logit & Probit) Comprehensive modelling including vintage, age and time effects Maximum likelihood estimation Significance testing Prediction Backtesting PDs via the validation() function: Discrimination and calibration AUROC Brier scores Binomial test Jeffreys prior test Calibration curve and outcome-fit diagram Tests for model stability

Day 2:

Duration	Topic
80 min	Machine Learning Concepts Statistical and machine learning thinking Workflow in machine learning Loss functions Train-test validation split Standardisation and scaling Bias-variance trade-off Cross-validation Hyper-parameters Supervised and unsupervised learning Classification and regression vs. dimensionality reduction
20 min	Break
80 min	Hands-on Excercises: Visualisation and Reporting Techiques Communication with non-Python users Excercise 1: Loan amortisation profiles for low and high risk borrowers Excercise 1: Validation framework Presentations in Jupyter Notebook and reveal.js User interfaces

Day 3:

Duration	Topic
80 min	Supervised Machine Learning for PDs - State of the Art Logistic regression K nearest neighbours Decision trees Bagging and boosting Random forests SGB, XGB and LGB
20 min	Break
80 min	Probabilities of Default (PD) – Crisis PDs Model stability, discrimination & calibration Through-the-cycle & point-in-time Predicting Financial Crises Asymptotic single risk factor model and Basel Capital Stress testing and Margin of Conservatism (MoC) CECL and IFRS 9 models for multi-period PDs

Day 4:

Duration	Topic
80 min	Hands-on Excercises: Hyperparmeter Tuning and Crossvalidation Excercise 1: Estimating a Neural Network Excercise 2: Tuning of Hyperparameters via Cross-validation: Network architecture and depth Learning rate Random number generator Activation Regularisation penalty ...
20 min	Break
80 min	Supervised Machine Learning for PDs - Future for Credit Risk Neural networks Deep learning Do it like Google: Introduction into Keras/Tenserflow

---

Back to Zoom Training