Analysis of univariate time series: description, modelling and forecasting. This module is aimed at the students who wish to gain a working knowledge of time series and forecasting methods using R.
Demonstrate understanding of the dynamic nature of the interrelationships between deterministic trend, seasonality and stochasticity within time series models.
Define and devise autoregressive integrated moving average (ARIMA) and seasonal ARIMA (SARIMA) models and evaluate their properties.
Apply Box-Jenkins methodology to select appropriate models for time series data, critically evaluate the merits of outcomes and create solutions for shortcomings.
Components of time series.
Trend analysis; moving average and exponential smoothing.
Holt-Winters methods.
Seasonal adjustment of time series, R-interface.
Probabilistic structure of time series; stationarity; autocorrelation functions.
Types of stochastic processes; white noise, random walk, autoregressive (AR), moving average (MA), and autoregressive integrated moving average (ARIMA) models.
Estimation, tests and model selection (Box-Jenkins methodology).
Seasonal autoregressive integrated moving average (SARIMA) models.
Assessment (if applicable):
Report based on time series with two parts: (1) to seasonally adjust the series (2) use Box-Jenkins methodology to find an adequate model for the data.