You are on page 1of 6

Exponential Smoothing

Perform exponential smoothing technique with alpha = 0.10, and compute for MAD, MSE, and MAPE
0.10 0.90
Actual
Period Forecast Absolute Error Error^2
Demand
January 300
February 356 300 56.00 3136
March 387 305.60 81.40 6626
April 381 313.74 67.26 4524
May 347 320.47 26.53 704
June 352 323.12 28.88 834
July 326.01
Total 260.07 15824.01
Mean Absolute Error (MAD) 52.01
Mean Standard Error (MSE) 3164.80
Mean Absolute Percent Error (MAPE)

Moving Average
Perform a 3-month moving average technique, compute for MAD, MSE, and MAPE
Actual
Period Forecast Absolute Error Error^2
Demand
January 300
February 356
March 387
April 381 347.67 33.33 1111.11
May 347 374.67 27.67 765.63
June 352 371.67 19.67 386.91
July 360
Total 80.67 2263.65
Mean Absolute Error (MAD) 26.89
Mean Standard Error (MSE) 754.55
Mean Absolute Percent Error (MAPE)

Time Series Regression


Perform a time series regression:
Actual
Period x xy x²
Demand (y)
January 1 300 300 1
February 2 356 712 4
March 3 387 1161 9
April 4 381 1524 16
May 5 347 1735 25
June 6 352 2112 36
Total 21 2123 7544 91
1. What is the slope (b)? 6.49
= (6 x 7544) - (21 x 2123)
= 45264 - 44583
= 681
= (6 x 91) - (21^2)
= 546 - 441
= 105
= 681 / 105 = 6.49
2.) What is the intercept (a-intercept)? 331.12
= 2123 - (6.49 x 21) / 6
= 2123 - (6.49 x 21) / 6
= 2123 - 136.29 / 6
= 1986.74/ 6
= 331.12
3. What is the forecasting equation? y= 331.12+6.49x
4. What is the forecast for September? 389.53
= 331.12 + (6.49 x 9)
= 331.12 + 58.41
389.53
for MAD, MSE, and MAPE

(Absolute Error Solution:


/AD)*100
February Forecast = 300, just use a naïve approach – meaning use the p
March Forecast= 356*0.10+300*0.90,
15.73 by subtracting 1 from 0.10.
21.03
February Absolute Error = 356-300 = 56
17.65 March Absolute Error = 387-305.60 = 81.40
7.65
February Error^2 = 56^2 = 3136
8.20 Mary Error^2 = 81.40^2 = 6626

April Percentage Error (last column) = (Absolute Error/Actual Demand)*


70.27
Note: Procedure is the same for all other months.

14.05 Mean Absolute Deviation (MAD) = Average of Absolute Error = 52.01


Mean Squared Error (MSE) = Average of Squared Error = 3164.80
Mean Absolute Percent Error (MAPE) = Average of Percent Error = 14.0

d MAPE
(Absolute Error Solution:
/AD)*100
April Forecast = (300+356+387)/3 = 347.67, the actual demand is divided
May Forecast = (356+387+381)/3 = 374.67
April Absolute Error = (381-347.67) = 33.33
8.75
May Absolute Error = (347-374.67) = 27.67
7.97
5.59 April Error^2 = 33.33^2 = 1111.11
Mary Error^2 = 27.67^2 = 765.63
22.31
Note: Procedure is the same for all other months.

7.44 Mean Absolute Deviation (MAD) = Average of Absolute Error = 26.89


Mean Squared Error (MSE) = Average of Squared Error = 754.55
Mean Absolute Percent Error (MAPE) = Average of Percent Error = 7.44%

Solution:

1. Slope (b)
= [n(total xy) – (total x)(total y)]/[n(total x^2)- (total x)^2]
= [6*(7544)-(21*2123)]/[(6*91)-(21)^2] = 6.49
Solution:

1. Slope (b)
= [n(total xy) – (total x)(total y)]/[n(total x^2)- (total x)^2]
= [6*(7544)-(21*2123)]/[(6*91)-(21)^2] = 6.49
where n is the number of periods

2. Intercept (a)
= [total y – (b*total x)]/n = 2123-(6.49x21)
= 331.12
where n is the number of periods
Where b is the computed slope

3. What is the forecasting equation


y = a + bx
y = 331.12+6.49x

4. Fore cast for September


y = 331.12+(6.49*9) = 389.53
h – meaning use the preceding period’s actual demand of 300

rror/Actual Demand)*100 = (56/356)*100 = 15.73%


.
solute Error = 52.01
Error = 3164.80
f Percent Error = 14.05%

tual demand is divided by 3 to get a 3-month moving average

olute Error = 26.89


Error = 754.55
Percent Error = 7.44%

You might also like