Course annotation

This course is intended to provide a comprehensive introduction to forecasting methods with an emphasis on application of statistical techniques to analysis and prediction of economic time series. Various relevant topics will be covered, such as: simple and multiple regression, basics of statistical hypothesis testing, model building and evaluation, criterions of forecast accuracy, autoregressive model, testing for stationarity of time series, testing for Granger causality by classical (1969) and Toda-Yamamoto (1995) methods, time series decomposition, seasonality and trend, exponential smoothing, autoregressive integrated moving average (ARIMA) model, Hyndman-Khandakar (2008) method for choosing ARIMA model order, time series intervention analysis by ARIMAX model. Practical examples having real-world relevance will be provided and students will be involved with hands-on experience of modelling and forecasting. Assignments will be solved using R packages and GRETL, but other econometrical software can also be used (for example: Excel, EasyReg, Statgraphics, RATS, GAUSS, JMulTi, SPSS Trends, S-Plus FinMetrics, Stata, Matlab, E-Views, SCA, SAS, OxMetrics).

Aim

The course is aimed at the students who need to have a basic knowledge of methods for time series analysis and obtain practical skills of forecasting, with a focus on economic applications.

Learning outcomes

• Apply various quantitative methods for economic forecasting.
• Construct econometric models and describe their structure.
• Evaluate accuracy of the forecast and be able to communicate forecasting results.
• Distinguish between short-term and long-term forecasting.
• Apply econometric software packages to produce forecasts and forecast confidence intervals.