Time Series Analysis

Autumn semester 2016

General information

Lecturer	Nicolai Meinshausen
Lectures	Wed 9-10 (HG D 7.1); Thu 10-12 (HG D 1.1)
Course catalogue data	>>

Course content

Statistical analysis and modeling of observations in temporal order, which exhibit dependence. Stationarity, trend estimation, seasonal decomposition, autocorrelations, spectral and wavelet analysis, ARIMA-, GARCH- and state space models. Implementations in the software R.

Announcements

September 1st, 2016:
Beginning of lecture: Wednesday, 20th September.

Course materials

Week	Topic
Week 1	Characteristics of time-series: stationarity, auto-correlation function, examples Simple time series models in R Book list Notes
Week 2	Characteristics of time-series: auto-correlation function and estimation Autocorrelation examples in R Notes
Week 3	Characteristics of time-series: transformations and trend estimation Trend and seasonality example in R Estimating trends for difference- and trend-stationary models Global temperature time-series 1880-2015 from NOAA
Week 4	Time-domain models Notes Simulation of AR(2) processes
Week 5	Invertible moving averages and ARMA models Notes Estimation of autocovariance from ARMA parameters and vice versa
Week 6	Linear forecasting and partial autocorrelations Linear forecasting of GDP data Function to estimate autocovariance for a given AR(p) model; via MA(infty) representation Swiss GDP data in comma-separated format (csv)
Week 7	Inference for ARMA models AIC-based model selection for ARMA processes Southern Oscillation Index data
Week 8	Spectral methods Notes Fourier transform examples
Week 9	State-space models Information about the exam Notes Location-guessing game in R Plotting function used in the R game
Week 10	State-space models Preliminary version of complete lecture nodes Document might get updated. The parts in small print were not discussed in class but can be optional additional reading.