Moving average

For time series without significant trend or seasonal variation

5 week moving
Week Demand average Absolute error
1 252922 265000
2 248559
3 253342
260000

Demand
4 249532
5 247693
255000
6 246618 250410 3792
7 261333 249149 12184
8 247447 251704 4257 250000
9 251492 250525 967
10 249713 250917 1204 245000
Demand
11 258563 251321 7242
12 249702 253710 4008 240000 Moving average (5 week)
13 262667 251383 11284
14 258829 254427 4402 235000
15 255631 255895 264 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
16 257078 Week
Mean absolute error 4960
Exponential smoothing
For time series without significant trend or seasonal variation

Week Demand Forecast Absolute error alfa = 0.24

1 252922 252936 265000
2 248559 252933 4374
3 253342 251883 1459
260000

Demand
4 249532 252233 2701
5 247693 251585 3892
255000
6 246618 250651 4033
7 261333 249683 11650
8 247447 252479 5032 250000
9 251492 251271 221
10 249713 251324 1611 245000 Demand
11 258563 250938 7625 Smoothed series
12 249702 252768 3066 240000
13 262667 252032 10635
14 258829 254584 4245 235000
15 255631 255603 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
16 255610 Week
Mean absolute error 4327
Simple linear regression for trend forecasting

Absolute 180
Week Demand Trend error
1 4.8 4.7 0.1 160
2 4 9.0 5.0
3 5.5 13.3 7.8 140
4 15.6 17.6 2.0
5 23.1 21.9 1.2

Demand
120
6 23.3 26.2 2.9
7 31.4 30.5 0.9
100
8 46 34.8 11.2
9 46.1 39.1 7.0
10 41.9 43.4 1.5 80
11 45.5 47.7 2.2
12 53.5 51.9 1.6 60
13 48.4 56.2 7.8 D
14 61.6 60.5 1.1 40
15 65.6 64.8 0.8
16 71.4 69.1 2.3 20
17 83.4 73.4 10.0
18 93.6 77.7 15.9
0
19 94.2 82.0 12.2
20 85.4 86.3 0.9
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
21 86.2 90.6 4.4 Week
22 89.9 94.9 5.0
23 89.2 99.2 10.0
24 99.1 103.5 4.4
25 100.3 107.8 7.5
26 111.7 112.1 0.4
27 108.2 116.4 8.2
28 115.5 120.7 5.2
29 119.2 125.0 5.8
30 125.2 129.3 4.1
31 136.3 133.6 2.7
32 146.8 137.9 8.9
33 146.1 142.2 3.9
34 151.4 146.5 4.9
35 150.9 150.7 0.2
36 155.0
37 159.3
38 163.6
39 167.9
Mean absolute error 4.9
 Demand Trend

17 19 21 23 25 27 29 31 33 35 37 39
Week
Additive model
Trend and seasonal variation considered

Moving Centered Absolute

Quarter Demand average moving average Trend Seasonal Forecast error Quarter Seasonal
1 1 8 18.96 -10.96 7.90 0.10 1 -11.06
2 2 13 19.49 -6.49 12.57 0.43 2 -6.92
3 3 23 19.50 19.75 20.01 2.99 23.90 0.90 3 3.89
4 4 34 20.00 20.63 20.53 13.47 35.57 1.57 4 15.04
1 5 10 21.25 21.25 21.06 -11.06 10.00 0.00 45.00
2 6 18 21.25 21.75 21.58 -3.58 14.67 3.33
3 7 23 22.25 22.50 22.11 0.89 26.00 3.00 40.00
4 8 38 22.75 22.13 22.63 15.37 37.67 0.33

Demand
35.00
1 9 12 21.50 22.63 23.15 -11.15 12.10 0.10
2 10 13 23.75 24.13 23.68 -10.68 16.76 3.76 30.00
3 11 32 24.50 24.20 7.80 28.10 3.90 25.00
4 12 41 24.72 16.28 39.76 1.24
20.00
1 13 25.25 14.19
2 14 25.77 18.86 15.00
3 15 26.30 30.19 Demand Trend
10.00
4 16 26.82 41.86
Forecast
Mean absolute error 1.56 5.00

0.00
Notes
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Quarter
ObservationS must be numbered 1, 2, 3, … (column B).
Seasonal variation is smoothed by using 4 quarter moving average.
First moving average 19,50 belongs between 2. quarter and 3. quarter.
Second moving average 20,00 belongs between 3. quarter and 4. quarter.
Average of 19,50 and 20,00 (centered moving average) belongs to 3. quarter.
Column E is a time series without seasonal variation.
Trend can be calculated based on column E.
Seasonal in column G can be calculated either by using additive or multiplicative model:
-in the multiplicative model seasonal is demand divided by trend
-in the additive model seasonal is demand minus trend.
Seasonal for each quarter (in column K) is the mean of quarter values in column G.
Final forecast is calculated either by using additive or multiplicative model:
-in the multiplicative model the final forecast is trend multiplied by season in column K
-in the additive model the final forecast is trend plus season in column K.
Multiplicative model
Trend and seasonal variation considered

Moving Centered Absolute

Quarter Demand average moving average Trend Seasonal Forecast error Quarter Seasonal
1 1 8 18.96 0.42 8.94 0.94 1 0.47
2 2 13 19.49 0.67 13.32 0.32 2 0.68
3 3 23 19.50 19.75 20.01 1.15 23.43 0.43 3 1.17
4 4 34 20.00 20.63 20.53 1.66 34.18 0.18 4 1.66
1 5 10 21.25 21.25 21.06 0.47 9.93 0.07 50.00
2 6 18 21.25 21.75 21.58 0.83 14.75 3.25
45.00
3 7 23 22.25 22.50 22.11 1.04 25.88 2.88
4 8 38 22.75 22.13 22.63 1.68 37.66 0.34 40.00

Demand
1 9 12 21.50 22.63 23.15 0.52 10.92 1.08 35.00
2 10 13 23.75 24.13 23.68 0.55 16.18 3.18
30.00
3 11 32 24.50 24.20 1.32 28.33 3.67
4 12 41 24.72 1.66 41.15 0.15 25.00
1 13 25.25 11.91 20.00
2 14 25.77 17.61
15.00
3 15 26.30 30.79 Demand Trend
4 16 26.82 44.64 10.00
Forecast
Absolute mean error 1.37 5.00
0.00
Notes
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Quarter
ObservationS must be numbered 1, 2, 3, … (column B).
Seasonal variation is smoothed by using 4 quarter moving average.
First moving average 19,50 belongs between 2. quarter and 3. quarter.
Second moving average 20,00 belongs between 3. quarter and 4. quarter.
Average of 19,50 and 20,00 (centered moving average) belongs to 3. quarter.
Column E is a time series without seasonal variation.
Trend can be calculated based on column E.
Seasonal in column G can be calculated either by using additive or multiplicative model:
-in the multiplicative model seasonal is demand divided by trend
-in the additive model seasonal is demand minus trend.
Seasonal for each quarter (in column K) is the mean of quarter values in column G.
Final forecast is calculated either by using additive or multiplicative model:
-in the multiplicative model the final forecast is trend multiplied by season in column K
-in the additive model the final forecast is trend plus season in column K.