S415 Machine Learning in Econometrics

Participation in this course is restricted and requires prior application. If you are interested in taking the class, please apply by sending an up-to-date version of your transcript of records to Prof. Joachim Grammig at  joachim.grammigspam prevention@uni-tuebingen.de using the subject line "S415: Application". The application deadline is April 18, 2021. All applicants will be informed on the outcome of their application by April 23, 2021.

Students taking part in the Advanced Time Series retake exam should comply with the application procedure described above. Your performance in the retake exam will be taken into account when considering your application.

This course is not suitable for bachelor students.

This module illustrates how machine learning techniques can be used in economic research and applications. It offers a thorough analysis of various tools in statistical learning and links them to econometric analysis. The course focuses on supervised machine learning techniques, such as: decision/regression trees, (logistic) regressions, naïve Bayes, local regressions, nearest neighbors, artificial neural networks, and support vector machines. The lecture also covers hyper-parameter tuning methods and different feature selection and regularization techniques. A virtual PC-lab class is an essential part of the module.

The lecture is accompanied by a practical class. During these sessions, students deepen their understanding by applying machine learning techniques to various problems using statistical software. The PC-lab class will be conducted live via Zoom. Details on the scheduling of these Zoom sessions will be announced in due time. All students are expected to be present during these sessions.

• Hastie/Tibshirani/Friedman: The Elements of Statistical Learning
• Bishop: Pattern Recognition and Machine Learning
• selected papers