Machine Learning for Economists (57750)

Spring 2024 @ Hebrew University of Jerusalem

Instructor: Itamar Caspi

Teaching assistant: Inbar Avni

🗓️ Second semester, 2024 ⏰ 16:30 - 19:15 🏨 Social Sciences 22205 ✍️ github.com/ml4econ 👪 Moodle Discussion Forum

---

Overview

This course encompasses a wide range of topics including data science, machine learning, and econometrics. Its primary objective is to acquaint students with the fundamental concepts of machine learning that can potentially enhance empirical economics. The course introduces popular supervised and unsupervised machine learning methods, while emphasizing the challenges and opportunities of incorporating these methods in empirical economics. Additionally, it highlights the relevance of machine learning to policy analysis and causal inference. The course utilizes real-world applications, empirical articles, and hands-on assignments to illustrate various topics.

Learning objectives

During the course, attendees will gain proficiency in the following:

1. Implementing data science best-practices in the context of empirical research within economics.

2. Acquiring comprehensive and pragmatic knowledge of the obstacles and prospects encountered in applied empirical work utilizing high-dimensional data.

3. Assimilating techniques and insights from the field of machine learning into applied empirical research conducted in economics.

Prework

It is anticipated that participants in the course will:

1. Possess their own computers with R, RStudio (Posit), Git, and GitHub Desktop installed.

2. Register for a GitHub and Kaggle account, which are available at no cost.

Schedule

The following schedule is subject to change based on class interests and time constraints. Our pace will be guided by class discussions, and we kindly request that you check this page regularly for updates.

Week	Topic
1	Course Overview
2	Basic ML Concepts
3	Reproducibility & ML Workflow
4	Regression and Regularization
5	Classification
6	Trees and Forests
7	Causal inference
8	High-Dimensional Counfounding Adjustment
9	High-Dimensional Heterogeneous Treatment Effects
10	Text Analysis
11	Large Language Models

Slides

Part I: Machine Learning

1. Course Overview (HTML) (PDF)

2. Basic Machine Learning Concepts (HTML) (PDF)

3. Reproducibility (HTML) (PDF)

4. ML Workflow (HTML) (PDF)

5. Regression and Regularization (HTML) (PDF)
   5.1 Prepare browser data
   5.2 Ridge and lasso simulation
   5.3 Ridge, lasso, PCR and PLS: A Tidymodels Workflow
   5.4 Shrinkage and selection intuition

6. Classification (HTML) (PDF)

7. Decision Trees and Random Forests (HTML) (PDF)

Part II: Causal Inference and ML

8. Causal Inference (HTML) (PDF)

9. High-Dimensional Confounding Adjustment (HTML) (PDF)

10. High-Dimensional Heterogeneous Treatment Effects (HTML) (PDF)

Part III: Unsupervised Learning and Language Models

11. Text as Data (HTML) (PDF)

12. Large Language Models (HTML) (PDF)

Projects

A. Kaggle competition (HTML) (PDF)

Readings

Can be found here.

People

• Itamar Caspi is the Head of the Monetary Analysis Unit in the Research Department of the Bank of Israel and an adjunct lecturer at the Hebrew University. His research interests include macroeconomics, monetary economics, and applied econometrics. He started his career in 2010 as an Economist at the Ministry of Finance. In 2012, he moved to the Bank of Israel and was promoted in 2018 to Senior Economist. He was also elected to represent the Bank as a Research Fellow at the Bank for International Settlements in Basel, Switzerland. Itamar holds a BA in Economics and Business Administration from Ben-Gurion University, an MA in Economics from the joint research program at Hebrew University and Tel-Aviv University, an MPA from Harvard Kennedy School, and a PhD in Economics from Bar-Ilan University.

---

This work is licensed under a Creative Commons Attribution 4.0 International License.