You will understand the theory behind models for time series data. These exhibit often rich dynamic structures, so data are dependent. Their analysis therefore requires techniques that are often different from cross-sectional settings. Theory is fully supported by extensive exercises and practical illustrations using the software R.
Course overview :
The course will focus on the analysis of a univariate time series.
The dynamics will be analysed by considering the autocorrelations of the series.
- We start with concepts of statistical regularities (weak and strong stationarity, asymptotic independence), then consider in detail the class of ARIMA models.
- Model estimation and model selection.
- Non-stationary models (deterministic and stochastic trends, structural breaks) ; unit roots.