The need to forecast or predict future values of economic time series arises frequently in many branches of applied economic and commercial work. It is, moreover, a topic which lends itself naturally to econometric and statistical treatment. The specific feature which distinguishes time series from other data is that the order in which the sample is recorded is of relevance. As a result of this, a substantial body of statistical methodology has developed. This unit provides an introduction to methods of time series analysis and forecasting. The material covered is primarily time domain methods designed for a single series and includes the building of linear time series models, the theory and practice of univariate forecasting and the use of regression methods for forecasting. Throughout the unit a balance between theory and practical application is maintained.
1x2hr lecture/week, 1x1hr lab/week
assignment (20%), group assignment (25%), Mid-semester test (20%) and 2.5hr Final exam (35%)
ECMT2010 or ECMT2110 or ECMT2030 or ECMT2130 or ECMT2160Prohibitions