1. The number of cases (in hundreds) of Merlot wine sold by Joe Cox Winery in Ar
ID: 352191 • Letter: 1
Question
1. The number of cases (in hundreds) of Merlot wine sold by Joe Cox Winery in Arrington, an unincorporated township off highway 96 east of Franklin between 2003 and 2015 are shown in the table below. Year Cases of Wine Sold (100s) 2003 380 2004 356 2005 410 2006 440 2007 358 2008 450 2009 410 2010 400 2011 362 2012 430 2013 390 2014 360 2015 415 a. Plot the data. Visually recommend what forecasting method you would use to forecast future sales. State your reasons. b. Suppose your boss has suggested that you pick one of the following methods (i) Simple Moving Average (MA) or Exponential Smoothing. Which would you choose and what will be the optima (best) parameters in that case. [Hint: specify 2-yr, 3-yr or 4-yr in the case of Moving Average; or the best value of , in case of exponential smoothing]. c. Also calculate the Tracking Signals for your forecasts using the best parameter of each method you arrived at above and comment on your results.
Explanation / Answer
When we plot the data, the following graph is obtained
We can see peaks and troughs. The sales shows cyclical nature as the interval between the peaks and the troughs are not constant. Hence, from visual analysis the recommended forecasting method would be cyclic ARMA Model
Given a choice between Simple Moving Average and Exponential Smoothing, the choice should be Simple Moving Average. Reason: The mean absolute percentage error for 2 year simple moving average is better than that of the best case of exponential smoothing (alpha = 0.09). The parameter to be chosen would be 2-yr simple moving average
Tracking signal for best parameter of Simple Moving average is as below (2 – year moving average) Tracking Signal = Running Sum of Forecasting Errors/(MAD)
Tracking Signal for Exponential Smoothing with alpha = 0.09 is as follows
Hence even by using the tracking signal it can be seen that 2-year moving average works better than exponential smoothing in this case.
Year Cases sold (in hundreds) 2 year moving average Error ABS Error MAD Running sum of Forecasting Errors Tracking Signal 2003 380 2004 356 368 -12 12 12.00 -12 -1.0 2005 410 383 27 27 19.50 15 0.8 2006 440 425 15 15 18.00 30 1.7 2007 358 399 -41 41 23.75 -11 -0.5 2008 450 404 46 46 28.20 35 1.2 2009 410 430 -20 20 26.83 15 0.6 2010 400 405 -5 5 23.71 10 0.4 2011 362 381 -19 19 23.13 -9 -0.4 2012 430 396 34 34 24.33 25 1.0 2013 390 410 -20 20 23.90 5 0.2 2014 360 375 -15 15 23.09 -10 -0.4 2015 415 387.5 27.5 27.5 23.46 17.5 0.7Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.