MAT3257

Forecasting models

Not offered in 1999

Dr Harmindar Nath

8 points - Second semester - 5 hours per week - Gippsland/Distance (even-numbered years) - Prerequisites: MAT1060 (MAT2236 is recommended) - Prohibitions: GAS3751, MAS3421, MAT2532, MAT3251

Objectives On the completion of this subject students will have developed the ability to identify decision situations requiring forecasts; understand the nature of the art of forecasting; distinguish a time series data set from a cross-sectional data set; recognise principal components underlying a time series; select and apply an appropriate forecasting technique from those covered in the subject to a given forecasting situation; model a time series using a quantitative forecasting technique; understand the limitations of a forecasting model; use a spreadsheet and a statistical package for modelling a time series.

Synopsis The subject aims to introduce the art of forecasting via a modelling approach. It is designed for the practioners of forecasting. The subject covers the nature and essentials of forecasting, introduction to time series modelling, importance of residual analysis; forecasting techniques- choice and applicability; the smoothing and decomposition methods of forecasting, causal methods of forecasting, ARMA modelling of stationary and non-stationary time series and regression models with correlated errors for forecasting. An overview of econometric models. Qualitative techniques of forecasting include Delphi method, Subjective probability method and Jury of Executive Opinion. The subject introduces the use of statistical software MINITAB and SPSS for windows to prepare and compare forecasts.

Assessment Two assignments (including a mini project): 60% - Examination: 40%

Prescribed texts

Makridakis S and others Forecasting: Methods and applications 3rd edn, Wiley, 1998

Recommended texts

Bowerman B and others Time series forecasting: Unified concepts and computer implementation 2nd edn, Duxbury, 1987
Farnum N and others Quantitative forecasting methods PWS-Kent, 1989

Back to the 1999 Science Handbook