The cafeteria supervisor at CarWash is interested in developing a forecast model for the number of meals to be prepared each day.
The following daily data from the cafeteria's register were collected. Using the data, the supervisor had already computed forecast values using a regression model.
However, s/he would like to assess the regression model's accuracy to a base model called "exponential smoothing."
WORKING AREA FOR THE PROBLEM
Exponential Regression
Day Meals Forecast Error Forecast Error
1 37 36.94
2 40 38.66
3 41 40.38
4 37 42.11
5 45 43.83
6 50 45.55
7 43 47.28
8 47 49.00
9 56 50.73
10 52 52.45
11 55 54.17
12 54 55.90
a) Your job is to compute forecasts using exponential smoothing technique with alpha of 0.3. Assume the initial forecast for day 1is 37. Show details for at least three calculations
b) Is the regression model a better fit to the data? Explain why. Use CFE and MAD as accuracy measurements.
c) Regardless of your decision in part (b), assume now that CarWash had been using the regression model in forecasting. What would you conclude about its performance?
Show details for updated MAD calculations, at least for three periods. Use action limits that correspond to 3? control limits, and plot the corresponding tracking signal values.
(HINT: 1sigma = 1.25 MAD, action limit is plus and minus and counts for how many MAD we have in specified number of sigma.)