Q : Regress widget sales on time and add the (a) line, (b) linear equation and (c) R^{2} (variance explained) to the graph you created in Q3. Please explain the meaning of each. Also, please characterize the trend if one appears to exist in the data (be specific).
Q: Take your understanding of seasonality from Q1 and use this understanding to employ a 3 month moving average to capture this seasonality effect. Please describe your forecasting method in words (8) and then implement this forecasting method (forecast through period 33) in Excel (10) (use "Data" tab, column G). Once you have completed implementing this method, assess its performance (RSFE, MAD, MSE, and MAPE using the appropriate columns I-AF).
For example, suppose that you observe an 8 period season (see Figure 1). You might develop a forecasting model for the NEXT SEASON that is based on observations from the PREVIOUS SEASON! For a more specific example, (see Figure 1) a 3-period moving average forecast for period 13 might appropriately be based on an average of periods (4,5 & 6)! To visualize this example, please see Figure 1.
Figure 1: Example of Seasonality (for example purposes only - - graph not based on actual exam data)
Q Consider both the forecasting method that you examined in Q5 in combination with the forecasting method that you examined in Q4. Combine these two methods to create a new forecasting method. Both describe this method in words (8) and implement this new forecasting method (forecast through period 33) in Excel (10) (use "Data" tab, column h). Once you have completed implementing this method, assess its performance (RSFE, MAD, MSE, MAPE using the appropriate columns I-AF) in comparison to the forecasts in Q2 and Q5 (1). Why is this (or is this not) a better forecasting method (2)?