Question

1. The Excel file “FoodBank” contains the number of pounds of food donated to a local food bank for each of the past 30 weeks. For planning purposes, the manager of the food bank would like a forecasting model to predict future donations. Create simple exponential smoothing models in Excel using α=0.2, α=0.5, and α=0.8. (Recall that damping factor = 1-α.)

Week	Donations
1	260
2	264
3	268
4	242
5	269
6	261
7	267
8	256
9	259
10	250
11	244
12	236
13	241
14	235
15	261
16	249
17	255
18	259
19	265
20	268
21	258
22	250
23	252
24	262
25	267
26	273
27	270
28	265
29	262
30	263

a. Which of the three models best fits the data set? How do you know that?

b. Using that model, predict the quantity of donations for week 31.

c. Import the data set into RapidMiner. Build a process to forecast the quantity of donations for future weeks using a Holt-Winters model to with the values of alpha, beta, and gamma all set to 0.5, and the period set to 1. Show a screenshot of the Process panel. You do not need to include the Parameters panel.

d. What quantity of donations does the Holt-Winters model predict for week 31?

Answer 1

Apply the below formula

Firstly we will calculate the simple smoothening model in excel

Using the excel

Go to Data>Data Analysis>Exponential smootheing>input range(Actual donations) fill value of damping factor>Ok

Or formula for forecasting

So table is given as below

Week	Donations (Actual)	Forecast (alpha(0.2)	Error(0.2)	Forecast (alpha(0.5)	Error(0.5)	Forecast (alpha(0.8)	Error(0.8)
1	260	#N/A	#N/A	#N/A	#N/A	#N/A	#N/A
2	264	260	#N/A	260	#N/A	260	#N/A
3	268	260.8	#N/A	262	#N/A	263.2	#N/A
4	242	262.24	#N/A	265	#N/A	267.04	#N/A
5	269	258.19	12.62	253.50	13.92	247.01	14.90
6	261	260.35	13.88	261.25	16.38	264.60	19.44
7	267	260.48	13.25	261.13	16.01	261.72	19.35
8	256	261.79	7.30	264.06	9.57	265.94	13.22
9	259	260.63	5.05	260.03	5.76	257.99	6.82
10	250	260.30	5.12	259.52	5.79	258.80	6.53
11	244	258.24	6.89	254.76	7.23	251.76	7.69
12	236	255.39	10.19	249.38	8.31	245.55	6.80
13	241	251.52	15.11	242.69	11.33	237.91	8.73
14	235	249.41	15.16	241.84	9.96	240.38	7.33
15	261	246.53	15.21	238.42	8.73	236.08	6.58
16	249	249.42	13.26	249.71	13.66	256.02	14.83
17	255	249.34	11.79	249.36	13.63	250.40	15.27
18	259	250.47	8.97	252.18	13.44	254.08	15.18
19	265	252.18	5.92	255.59	5.13	258.02	5.61
20	268	254.74	9.47	260.29	7.46	263.60	5.60
21	258	257.39	11.73	264.15	8.05	267.12	5.55
22	250	257.51	10.65	261.07	7.87	259.82	7.10
23	252	256.01	8.81	255.54	8.56	251.96	8.15
24	262	255.21	4.93	253.77	7.59	251.99	7.74
25	267	256.57	6.29	257.88	8.22	260.00	8.10
26	273	258.65	7.55	262.44	7.38	265.60	7.05
27	270	261.52	10.97	267.72	9.35	271.52	8.24
28	265	263.22	11.35	268.86	8.16	270.30	5.95
29	262	263.57	9.68	266.93	6.62	266.06	5.33
30	263	263.26	5.08	264.47	3.85	262.81	3.96

Let us find the values of Average rediuals(error)

Whose residuals(error) will be least model will be best

Model 1 when alpha =0.2

Residual=Average(Actual-Forecast)=9.86

Model 1 when alpha =0.5

Residual=Average(Actual-Forecast)=9.31

Model 1 when alpha =0.8

Residual=Average(Actual-Forecast)=9.27

The model with alpha =0.8 is best model because error is least

b)

We have to predict or Forecasted value for donations for week 31

c)