|
Lehrende: Prof. Dr. Melanie Krause Ph.D.
Veranstaltungsart:
Interaktive Lehrveranstaltung
Anzeige im Stundenplan:
22-3.e72
Semesterwochenstunden:
3
Credits:
6,0
Unterrichtssprache:
Englisch
Min. | Max. Teilnehmerzahl:
- | 72
Kommentare/ Inhalte:
Machine learning (ML) has become a buzzword of our time. With the amounts of data available growing exponentially, there is an increased demand to make use of all this information. In this course we take a look at the main methods of supervised and unsupervised learning from an economist's point of view. In particular, we will (i) learn about the most widely-used ML techniques such as variance-reduction techniques, neural networks and random forests, (ii) apply them to real-word data, (iii) discuss how ML-based prediction works differently from econometric inference.
Lernziel:
Course participants will get to know the most important ML methods which are currently in use. The focus is on how ML methods can complement the statistical toolkit of an economist for research and applications. Course participants will learn how to apply a range of ML methods in Python, thus becoming well-equipped to conduct their own empirical project.
Vorgehen:
There will be weekly three-hour lectures including work-through examples and applications. These lectures will be recorded and made available to the course participants via OpenOLAT. The OpenOLAT page will also contain the slides, code, data as well as a forum for discussions. The asynchronous lectures are complemented by weekly Zoom sessions for discussing the material, asking questions and interacting with other course participants.
Literatur:
Main Course Books and Papers:
• Hastie, Tibsharini and Friedman (2017), “The Elements of Statistical Learning - Data Mining, Inference, and Prediction“, Second Edition.
• Mueller and Guido (2017), “Introduction to Machine Learning with Python“, First Edition.
• Athey and Imbens (2019), “Machine Learning Methods Economists Should Know About“, Working Paper.
No prior knowledge of machine learning and programming in Python is required. However, all course participants should have a solid background in econometrics.
Zusätzliche Hinweise zu Prüfungen:
In order to pass the course, students must (a) pass an online-presentation and (b) complete a short research project involving ML methods learned in the course. Participants are encouraged to suggest their own research question and data set, explaining why ML is particularly suited in this situation.
The presentation will make up 30% of the final grade and the project 70%.
Presentation dates: 24, 25 and 26 February 2021 from 1:00 pm until 6:00 pm
|
