ECON30640

Learning Outcomes:

On successful completion of the module, students should be able to:

Formulate regression, classification, forecasting, clustering, and dimension-reduction tasks using appropriate outcomes, predictors, and loss functions.
Explain the central ideas underlying regularisation, cross-validation, and the bias–variance trade-off,
Apply popular machine learning methods, including tree-based methods, ensemble methods, and neural networks.
Select and compare predictive models using out-of-sample evaluation procedures and performance measures.
Process a range of data types for machine learning methods in social science, e.g. text data, time series data, high dimensional data.

Indicative Module Content:

Topic 1: Introduction and overview: research questions, machine learning tasks, types of data.
Topic 2: (Linear and Nonlinear) Regression: predicting continuous variables. Regularization, variable selection, boosting, bagging, tree-based methods, and neural networks.
Topic 3: Classification: predicting categorical variables. Regularized logistic regression, trees, and neural networks for classification. Loss functions and calibration. Ordered and unordered categorical variables.
Topic 4: Forecasting/nowcasting: predicting variables in a time series context with machine learning. Autoregression, forecast evaluation, macroeconomic nowcasting.
Topic 5: Unsupervised Learning, dimension reduction: principal components analysis, partial least squares, and nonlinear dimension reduction methods.
Topic 6: Text as Data: co-occurrence of words, latent Dirichlet allocation, embeddings, sentiment analysis, and customer reviews.

The United Nations identified seventeen Sustainable Development Goals (SDGs) as core to the 2030 Agenda for Sustainable Development, and UCD contributes in general to SDG 4 Quality Education. Further SDGs explored within this module if relevant are listed below. A scale of 1 - 5 indicates the extent to which the SDG is covered.


 

Student Effort Hours:

Approaches to Teaching and Learning:

There will be two one-hour lectures per week. Lecture slides will be made available on Brightspace. Ungraded practice problems will be made available.
Students may use AI for independent learning.
Both the mid-term and final exam will be closed book exams, with no access to AI.

Requirements, Exclusions and Recommendations

Learning Requirements:

Students are required to have the following two modules:

(1) an introductory statistics module such as ECON10800 Stats Methods for Economists, STAT10060 Statistical Modelling, MIS10090 Data Analysis for Decision Makers, or an equivalent module.

(2) introductory mathematics modules such as: ECON10790 Mathematics for Economists, MST10010 Calculus I, MST10030 Linear Algebra I, MATH10120 Linear Algebra with Applications to Economics, or equivalent modules.

Learning Recommendations:

An intermediate mathematics or statistics module such as ECON 20040 Statistics for Economists, MATH20330 Optimisation for Economics, MST20050 Linear Algebra II, STAT20200 Probability, STAT20240 Predictive Analytics, or an equivalent module

An advanced module such as: ECON30130 Econometrics, ECON30540 Advanced Econometrics: Time Series, STAT30010 Time Series, or an equivalent module.

Module Requisites and Incompatibles

Not applicable to this module.
 

Assessment Strategy

As part of UCD's student support, under the Additional Consideration policy, extensions may be available for the following assessments in the module: Assignment (including essay/poster), Portfolio, Reflective Assignment, Report(s), and Individual Project.

 

Please see Student Jargon Buster for more information about remediation types and timing. 

Feedback Strategy/Strategies

• Feedback individually to students, post-assessment
• Group/class feedback, post-assessment

How will my Feedback be Delivered?

Grades for assessments will be made available to students post-assessment. Detailed assessment solutions will also be made available to students.

The main textbook for most, but not all, topics will be:
ISL. James, Gareth, Daniela Witten, Trevor Hastie, and Robert Tibshirani. An introduction to statistical learning: with applications in R. Second edition. New York: springer, 2021.
Free digital copy available at: https://www.statlearning.com/
Data and code: https://www.statlearning.com/resources-second-edition

The above text does not focus on social science applications. For social science and business applications, the following textbooks will often be referenced:
BDS. Taddy, Matt. Business Data Science: Combining Machine Learning and Economics to Optimize, Automate, and Accelerate Business Decisions. New York: McGraw Hill, 2019.
Code examples are available at: https://github.com/TaddyLab/BDS https://github.com/TaddyLab/MBAcourse . Available from UCD library as an e-book and there are some print copies.
A less technical alternative is:
MBA. Taddy, Matt, Leslie Hendrix, and Matthew C. Harding. Modern business analytics: practical data science for decision-making. McGraw Hill, 2023.
Not available as an e-book from UCD library, but there are some print copies.

A detailed reading list with topic-specific reading suggestions, including references to other textbooks, will be made available on Brightspace.

Timetabling information is displayed only for guidance purposes, relates to the current Academic Year only and is subject to change.