Australian rail freight forecasts
2007 – 2008 and 2015
Executive summary
In this report, both the forecast of Australian rail freight in the year 2007 and 2015 are analyzed, in which the exponential smoothing method and regression method are used. This report can give a general picture of the rail freight in the next few years.
Rail freight forecasting for year 2007 – 2008
2.1 Overview of exponential smoothing
To forecast the rail freight in the period of 2007-2008, exponential smooth method should be used.
Exponential smoothing is also called ‘exponentially weighted moving average’. It weights the more recent periods of data more heavily in the forecast, and the weight for the older periods will decrease at an exponential rate.
The simple formula of exponential smoothing is:
F_(t+1)=aA_t+(1-a)F_t
where F_t=forecast for period t, A_t=available data for period t, and 0 < a < 1.
2.2 Data collection
From the formula F_(t+1)=aA_t+(1-a)F_t, if it is to forecast the rail freight in 2007, then the data of year 2005 and 2006 is needed.
According to Australasian Railway Association’s Industry Report 2006 and 2007, the rail freight respectively is 189.0btk in 2005-2006 and 198.7btk in 2006-2007. (see Table 1)
Table 1: NET TONNE-KILOMETRES (BILLION)
2.3 Estimate alpha
When a=0.9, the standard error s = 12(btk), when a=0.7, s = 15(btk), when a=0.5, s = 20 (btk) in the year 2006, so it is reasonable to estimate a=0.9. The Mean Absolute Deviation (MAD) =
2.4 Forecasting
When a=0.9, the forecast of rail freight in 2007-2008 is 197.9 BTK. With the MAD of 5.6 BTK, the forecast freight can be a number between 192.3 BTK and 203.5 BTK.
Table 2: Forecasting rail freight by exponential smoothing method ...