Description: Time-series forecasting is crucial for decision making processes in regard to behavioral dynamics in many domains. For example, balancing the supply and demand of limited resources requires accurate and timely estimates of future consumption and production, and may further include information about the forecasting uncertainty and flexibility.
In this study group, the students will become acquainted with both fundamental theories and concepts of statistical time-series analysis as well as advanced topics in this area of research. Modeling of both univariate and multivariate time series will be discussed. Topics can include standard regression and exponential smoothing techniques, spectral methods and filtering, switching models, time-series clustering, and real-time analysis for fast streaming data. Advanced methods for discovering structural breaks, abnormalities, and impact of planned events may also be considered.

Student presentations, discussions, and overview lectures.

Some computational and mathematical maturity is expected. The students attending this study group are, in particular, expected to have basic working knowledge of statistics and probability theory.

Learning objectives:
Heavy emphasis will be given to fundamental concepts, with the goal to provide starting points for students that wish to include more in-depth time-series analysis in their further work.

Organizer: Bo Thiesson, Associate Professor,

Lecturers: Thomas Dyhre Nielsen, Associate Professor, and Bo Thiesson, Associate Professor,


Time: 31 March, 19+28 April and 12+26 May, 2014

Place: Aalborg University, Department of Computer Science, Selma Lagerlöfs Vej 300

Zip code: 9220

City: Aalborg

Number of seats:

Deadline: 10 March, 2014