Course material and links for Promotionskolleg module "Machine learning in applied economics"
Kathy Baylis - University of California, Santa Barbara, USA
Thomas Heckelei - University of Bonn, Germany
Hugo Storm - DHL and University of Bonn, Germany
- Day 1 slides
- Day 1 video part I - Intro to ML with OLS
- Day 1 video part II - Overfitting and train/test split
- Day 1 video part III - Penalized Regression
- Day 1 jupyter notebook for lecture and lab
- Day 2 slides
- Day 2 video part I - Intro to trees
- Day 2 video part II - Random Forest and Boosted trees
- Day 2 video part III - Intro to interpreting ML Models
- Day 2 video part IV - Visualizing marginal effects
- Day 2 jupyter notebook for lecture and lab
- Day 3a slides - Interpretation part II, Shapley values and other approaches
- Day 3b slides - Neural Networks (also include part of day 4 slides)
- Day 3 video part I - Shap-values
- Day 3 video part II - Other approaches to interpreting ML Models
- Day 3 video part III - Intro Neural Networks and Autoencoder
- Day 3-4 jupyter notebook for lecture and lab
- Day 4a slides - Neural Networks (same as day 3)
- Day 4b slides - ML and causal analysis (also include part of day 5 slides)
- Day 4 video part I - Types of NN
- Day 4 video part II - NN-applications
- Day 4 video part III - Review of causal identification issues
- Day 4 video part IV - True model selection with LASSO
- Day 4-5 jupyter notebook for lecture and lab