Time and place

Lecture: Do, 10-12h, SR 127, Ernst-Zermelo-Str. 1
Tutorial: 2 hours, date to be determined

Teacher: Rainer Dahlhaus

From a narrow perspective, time series analysis is the statistical study of the properties of stochastic processes in discrete time. In this lecture, we will take a broader view: First we will examine the important probabilistic properties of stationary processes, including strong laws of large numbers (based on the Ergodic theorem) and various versions of the central limit theorem (for processes with strong dependence, even the rate of convergence can change). Another exciting topic is the relation between stationary processes and Fourier analysis based on the Cramér-representation, in which a stationary process is represented as a Fourier-integral of a stochastic process in continuous time (such as the Brownian motion). This later leads, on the statistical side, to a quasi-maximum likelihood method in the frequency domain. Furthermore, we investigate linear and nonlinear time series models, the prediction of time series, linear filters, linear state space models, model selection, maximum likelihood and quasi maximum likelihood methods, the Toeplitz-theory for quadratic forms of stationary processes. Finally, we provide an outlook on locally stationary processes, where the process can be locally apprximated by stationary processes.

Elementary Probability Theory I (Stochastik I) and Probability Theory (Wahrscheinlichkeitstheorie)

Elective (Option Area) (2HfB21)
Compulsory Elective in Mathematics (BSc21)
Supplementary Module in Mathematics (MEd18)
Applied Mathematics (MSc14)
Mathematics (MSc14)
Concentration Module (MSc14)
Elective (MSc14)
Elective in Data (MScData24)