The course is an introduction to time series econometrics for second-cycle studies and treats basic themes in modern time series analysis. A student who has taken the course should: * have a solid knowledge about basic themes in modern time series analysis * know and be able to use concepts and notation that is frequently used in time series analysis * know and be able to use different probabilistic results for serially dependent observations * be familiar with different methods to estimate time series models * be able to choose on appropriate model and estimation method for a given time series * be able to interpret the results of an fitted model * be aware of limitations and possible sources of errors in the analysis
Difference equations. White noise, stationarity and ergodicity. Stationary ARMA processes: The Box-Jenkins approach. Prediction: Wold's theorem, test for predictive accuracy. Vector autoregressive models. Maximum likelihood estimation. Asymptotic theory for serially dependent vector autoregressive processes. Bayesian analysis. State-space model and the Kalman filter. Models of nonstationary time series: unit root and deterministic time trends. Cointegraton.
Teaching is given in the form of lectures and tutorial classes.
The examination takes place through a written examination at the end of the course and compulsory written assignments. The grading scales are: failed, passed and passed with distinction.
The course is included in the Master's programme in statistics.