Forecasts and actual sales of Fidget spinners in a convenient store are as follo
ID: 366318 • Letter: F
Question
Forecasts and actual sales of Fidget spinners in a convenient store are as follows:
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
170
229
192
241
238
210
225
179
What is the forecast for November using a two-period moving average? [5 points]
What is the forecast for November using a three-period moving average? [5 points]
Calculate the forecasted demand using 2-period and 3-period moving average methods. What are the Mean Squared Errors (MSE) for the two- and three-period moving average models? [5 points]
Based on the MSE, which number of periods for the moving average should be used for the model? (2 period or 3 period)? Describe why. [5 points]
What are the forecasts for November using single exponential smoothing with =0.1 and =0.8. [5 points]
Calculate the forecasted demand using single exponential smoothing with =0.1 and =0.8. What are the MSE for two methods? [5 points]
What is the best single exponential smoothing model by evaluating MSE? Describe why. [5 points]
Plot the Sales and Forecast data for (f). Provide insides about the time series. [5 points]
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
170
229
192
241
238
210
225
179
Explanation / Answer
There are too many sub questions. I will help you with the first 4.
1) Forecasted demand for November using a 2 period moving average:
Note the table above. What I have done is, I have calculated the 2 period moving averages till November.
Note that we can calculate the 2 period moving average only if we have at least 2 data points in history. Hence, we can start our actual forecast only from May as only then do we have two data points.
The 2 period moving average for May = (Demand in April + Demand in March)/2 = (170+229)/2 = 199.5
Similarly, the 2 period moving average for Jun will consider the Demand in May and the Demand in April. Add those two and divide by 2
Thus, you see from the table above that the forecast for November is 202 using the 2 period moving average
What is the forecast for November using a three-period moving average? [5 points]
Using the 3 period moving average, it is very simple - Just consider the last 3 periods instead of the last 2 periods.
Note that our forecast in this case starts from June, as we need at least 3 data points.
This time, we consider the last 3 months data, add them and divide by 3 to calculate the forecast. I trust you can do this one.
From the table above, you can see that the forecast for November using 3 point average is 204.67
Calculate the forecasted demand using 2-period and 3-period moving average methods. What are the Mean Squared Errors (MSE) for the two- and three-period moving average models?
To calculate the MSE, simply subtract the forecast from the actual, Square it, and add all these terms. Now, simply take an average of all these squared terms.
Do this for both the 2 point and the 3 point results you have calculated. The answers you can see above.
The MSE is an indication of the error - higher this number, more inaccurate is the method.
For the given data, you can clearly see that the 2 point method seems more suited for this data than the 3 point method due to a lower MSE
Month Actuals 2 month average Mar 170 Apr 229 May 192 199.5 Jun 241 210.5 Jul 238 216.5 Aug 210 239.5 Sep 225 224 Oct 179 217.5 November 202