If i have 50 randomly generated numbers and I want to apply the following forecasting techniques on them; Simple moving average, Exponential smoothing and trend projection with regression, which would be the best technique between them and why will I choose this one?
Week | Demand |
1 | 688 |
2 | 681 |
3 | 643 |
4 | 667 |
5 | 642 |
6 | 683 |
7 | 758 |
8 | 791 |
9 | 751 |
10 | 682 |
11 | 609 |
12 | 635 |
13 | 647 |
14 | 796 |
15 | 676 |
16 | 617 |
17 | 704 |
18 | 722 |
19 | 764 |
20 | 726 |
21 | 692 |
22 | 708 |
23 | 674 |
24 | 606 |
25 | 640 |
26 | 785 |
27 | 717 |
28 | 684 |
29 | 665 |
30 | 713 |
31 | 711 |
32 | 666 |
33 | 779 |
34 | 756 |
35 | 691 |
36 | 641 |
37 | 617 |
38 | 729 |
39 | 768 |
40 | 621 |
41 | 771 |
42 | 760 |
43 | 684 |
44 | 683 |
45 | 603 |
46 | 665 |
47 | 720 |
48 | 692 |
49 | 704 |
50 | 647 |
Note that the numbers are generated randomly i.e. from a uniform distribution. So, it is expected that the resultant series will not have any trend or seasonality. It will only be a random variation or white noise.
Therefore, there is no need to go for regression as there will be no trend.
Among simple exponential smoothing and smile moving average, a simple moving average method will be the best method in this case because we do not have to account for trend or seasonal pattern and the moving average is the simpler among the two.
Get Answers For Free
Most questions answered within 1 hours.