PROGRAMME
Monday, July 13
Intro, prediction versus estimation, overfitting and regularization -
Kathy Baylis, Thomas Heckelei
09:00-10:30 Recorded Lecture 1a: Introduction to ML basics
10:30-12:00 Recorded Lecture 1b: Penalized regressions
11:45-12:00 Live Session - Welcome to Giovanni Anania' Summer School
12:00-13:00 Live Session - Questions and Answers on the recorded material
13:00-14:30 lunch break
14:30-15:30 Live Session - Lab 1a: Introduction to Jupyter
Notebooks and summary stats using Python
15:30-16:00 break
16:00-17:00 Live Session -
Lab 1b: Comparing OLS, LASSO, Ridge
and ElasticNet
17:00-18:00 Live Session -
Unstructured lab time
Tuesday, July 14
Trees, forests and how to not get lost (Interpretability vs complexity)
- Kathy Baylis, Thomas Heckelei
09:00-10:30 Recorded Lecture 2a: Tree-based methods
10:30-12:00 Recorded Lecture 2b: Interpretation
12:00-13:00 Live Session - Questions and Answers on the recorded material
13:00-14:30 lunch break
14:30-15:30 Live Session - Lab 2a: prediction using tree-based
methods
15:30-16:00 break
16:00-17:00 Live Session - Lab 2b: Interpretation using Effects
Plots (PDP, ICE) and Shapley Values
17:00-18:00 Live Session - Unstructured lab time
Wednesday, July 15
Neural networks - Gianluigi Greco
09:00-10:00 Recorded Lecture 3a: Introduction to Neural Networks
10:00-11:00 Recorded Lecture 3b: Neural networks for regression,
binary classification, and multiclass classification
11:00-12:00 Recorded Lecture 3c: Analysis of time series, advanced
network architectures
12:00-13:00 Live Session - Questions and Answers on the recorded material
13:00-14:30 lunch break
14:30-15:30 Live Session - Lab 3a: Examples of Neural Networks
on real-world data
15:30-16:00 break
16:00-17:00 Live Session - Lab 3b: Examples of Neural Networks
on real-world data
17:00-18:00 Live Session - Unstructured lab time
Thursday, July 16
Machine Learning for causal analysis - Kathy Baylis, Thomas Heckelei
09:00-10:30 Recorded Lecture 4a: Model selection, Matching and Doubly
robust regression
10:30-12:00 Recorded Lecture 4b: Overview of methods for causal ID
12:00-13:00 Live Session - Questions and Answers on the recorded material
13:00-14:30 lunch break
14:30-15:30 Live Session - Lab 4a: LASSO for model selection
and PSM
15:30-16:00 break
16:00-17:00 Live Session - Lab 4b: Double ML using LASSO as
selection
17:00-18:00 Live Session - Unstructured lab time
Friday, July 17
Superlearning Machine and Stata-Python Integration - Giovanni Cerulli
09:00-10:30 Online Lecture 5: The ontology and practice of Machine
Learning: an overview
10:30-12:00 Personal study by students on the lecture material
12:00-13:00 Live Session - Questions and Answers on the recorded material
13:00-14:30 lunch break
14:30-15:30 Live Session -
Lab 5a: The superlearning machine
for predicting economic outcomes
15:30-16:00 break
16:00-17:00 Live Session -
Lab 5b: The Stata/Python
integration for Machine Learning purposes
