Deep Learning in Energy

Module: Deep Learning in Energy (6 Credits)
Name in diploma supplement	Deep Learning in Energy
Responsible	Prof. Dr. Florian Ziel
Admission criteria	See exam regulations.
Workload	180 hours of student workload, in detail: Attendance: 60 hours Preparation, follow up: 80 hours Exam preparation: 40 hours
Duration	The module takes 1 semester(s).
Qualification Targets	The students have an advanced understanding of electricity markets and systems understand deep learning based modeling methods for energy markets and systems can apply learning and forecasting algorithms to real data using deep learning software able to interpret and to visualize the results
Module Exam	Equally weighted average of a group project and a presentation (usually about 20 minutes).
Usage in different degree programs	BWL EaF Master > Wahlpflichtbereich > 1.-3. Sem, Elective ECMX Master > Wahlpflichtbereich > ME6 Applied Econometrics > 1.-3. Sem, Elective MuU Master > Wahlpflichtbereich I > Wahlpflichtbereich I A.: Methodologie und allgemeine Theorien zur Untersuchung von Märkten und Unternehmen > 1.-3. Sem, Elective VWL Master > Wahlpflichtbereich II > 1.-3. Sem, Elective
Elements	Lecture Deep Learning in Energy (3 Credits) Exercise Deep Learning in Energy (3 Credits)
Module: Deep Learning in Energy (WIWI‑M0967)

Lecture: Deep Learning in Energy (3 Credits)
Name in diploma supplement	Deep Learning in Energy
Organisational Unit	Lehrstuhl für Data Science in Energy and Environment
Lecturers	Prof. Dr. Florian Ziel
Hours per week	2	Language	English
Cycle	irregular	Participants at most	24
Preliminary knowledge Good knowledge of linear models as tought in Econometrics of Electricity Markets and R or python knowledge
Abstract The objective of the lecture is to provide a basic understanding of energy markets and systems such as deep learning based modeling methods with a focus on feed forward neural network and recurrent neural networks. The aim of this course is to understand and apply deep learning algorithms to real data using the pytorch library, to interpret and to visualize the results.
Contents Introduction to electricity markets Overview of different non-linear model approaches Advanced forcasting study design, (hyperparmeter) optimization/learning, evaluation and ensembling Feed forward and recurrent neural networks and in detail
Literature The relevant material will be given during the course. Suggested reading: Weron, Rafał. "Electricity price forecasting: A review of the state-of-the-art with a look into the future." International Journal of Forecasting 30.4 (2014): 1030-1081. Petropoulos, F., Apiletti, D., Assimakopoulos, V., Babai, M. Z., Barrow, D. K., Taieb, S. B., ... & Ziel, F. (2022). Forecasting: theory and practice. International Journal of Forecasting, 38(3), 705-871. Marcjasz, G., Narajewski, M., Weron, R., & Ziel, F. (2023). Distributional neural networks for electricity price forecasting. Energy Economics, 125, 106843. Goodfellow, I. (2016). Deep learning.
Teaching concept Lecture. The studied modeling an forecasting methods are applied on real data using pytorch.
Lecture: Deep Learning in Energy (WIWI‑C1265)

Exercise: Deep Learning in Energy (3 Credits)
Name in diploma supplement	Deep Learning in Energy
Organisational Unit	Lehrstuhl für Data Science in Energy and Environment
Lecturers	Prof. Dr. Florian Ziel
Hours per week	2	Language	English
Cycle	irregular	Participants at most	24
Preliminary knowledge See Lecture
Contents See Lecture
Literature See Lecture
Teaching concept Tutorials. The students apply the learned methods in a own real data project in python utilizing pytorch.
Exercise: Deep Learning in Energy (WIWI‑C1266)