Machine Learning for Economists
Module Code: ECP77453
- ECTS Credit: 5
- Mandatory/Optional: Mandatory
- Semester/Term Taught: Semester 1
- Module Coordinator: Professor Niamh Wylie
Aims of Module
The module is designed to introduce students to the field of Machine Learning and its broad applications in Economics for prediction and classification tasks. Students will explore a range of models and techniques for data analysis and Natural Language Processing(NLP). The algorithms will include Binary Logistic Regression, Naïve Bayes, Support Vector Machines, Decision Trees, Random Forest, Gradient Boost as well as an introduction to Deep Learning and Neural Networks for Large Language Models (LLMs). The module will be both applied and theoretical in nature. Students will be expected to complete a weekly exercise for continuous assessment and a final individual assignment.
Learning Outcomes
On completion of the module, students will be able to:
- Differentiate between traditional econometric and machine learning (ML) methods.
- Understand and apply an array of supervised and unsupervised ML algorithms to analyse economic, financial and textual data.
- Evaluate the accuracy of various ML models for classification and prediction purposes
- Demonstrate a working proficiency in Python, and how to source data.
- Describe and appraise how ML is used in Economics - its applications, benefits and limitations.
Module Content
| Content | Indicative Number of Hours |
|---|---|
Lecturing hours |
10 |
Tutorial hours |
5 |
Preparation for lectures |
5 |
Individual continuous assessment |
20 |
Reading of assigned materials and active reflection on lecture and course content |
30 |
Individual assignment |
20 |
Total |
90 |
This module is structured around a series of lectures and tutorials. Lecture slides will be provided in advance of lectures. Lectures will be delivered over 5 weekly 2-hour blocks and a weekly 1-hour tutorial to practically apply the learnings. Python programming will be used in class and students will be expected to submit assessments using Python also. Reference will be made to Journal articles to strengthen the links from theory to practice. Useful resources such as lecture notes, data sets and readings will be available on Blackboard.
| Session | Dates | Content |
|---|---|---|
1 |
16 Sep 24 | Introduction to Machine Learning (ML) and Python |
2 |
23 Sep 24 | ML for Classification and Prediction |
3 |
30 Sep 24 | Supervised and Unsupervised ML models |
4 |
7 Oct 24 | Deep Learning, Neural Networks |
5 |
14 Oct 24 | Applications of ML in Economics |
Module Pre-Requisites
None
Module Co-Requisites
- ECP 77403 Introduction to Statistics and Regression Analysis
- ECP 77413 Introduction to Big Data for Economics
- ECP 77421 Microeconometrics
Assessment Details
Assessment for this course will be as follows:
- Continuous Assessment - 50%
- Individual Assignement - 50%
Module Preparation
Students should be comfortable with basic statistics and probability theory. Students will have to use Python for their continuous assessment so having a basic understanding would be useful, but not essential, prior to class commencing.
The following readings are recommended:
- Breiman, L. (2001) ‘Statistical modeling: The two cultures’, Statistical Science, 16(3), pp. 199–215. doi: 10.1214/ss/1009213726.
- Athey, S. 2019. 21. The Impact of Machine Learning on Economics. In: Agrawal, A., Gans, J. and Goldfarb, A. ed. The Economics of Artificial Intelligence: An Agenda. Chicago: University of Chicago Press, pp. 507-552. https://doi.org/10.7208/9780226613475-023
Course Communication
Typically via email and during office hours, and I will be available for feedback on continuous assessment and questions on the assignment. This will be communicated to students in due course. Any changes to the timetable or lectures will be announced on Blackboard.
Please note that all course related email communication must be sent from your official TCD email address. Emails sent from other addresses will not be attended to.