Forecasting
Forecasting
Forecasting
Forecasting
Characteristics of Forecasts:
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
The Delphi method attempts to eliminate some of the
inherent shortcomings of group dynamics, in
which the personalities of some group members
overshadow those of other members.
The method requires a group of experts to express
their opinions, preferably by individual sample
survey. The opinions are then complied and a
summary of the result is returned to the experts.
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
That is,
Y = + X + X + ................. + X
0
1 1
2 2
n n
for some constants (1....., n).
Forecasting
Forecasting
Forecasting
Forecasting
Yt = 385.7 1,878Xt-1 ,
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Ft =
some set of weights a1,
afor
n Dt n
n 1
a2,......
Most of the time series methods discussed in
this chapter are distinguished only by the
choice of weights.
Forecasting
Forecasting
Forecasting
n
MAD =
1/n
| e |
i 1
MSE
1/n
2
|
e
i|
i 1
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Week
P1
O1
E1
E12
|E1/O1|
P2
O2
E2
E22
|E2/O2|
92
88
16
0.04545
95
91
25
0.05495
87
88
0.01136
89
89
95
97
0.02062
92
90
0.02222
90
83
49
0.08434
93
90
0.03333
88
91
0.03297
90
86
16
0.04651
93
93
85
89
16
0.04494
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
MAPE2 = .0336.
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Engine
Quarter
1
Failures
200
MA(3)
Error
250
175
MA(6)
186
208.33
22.33
225
203.67
-21.33
285
195.33
-89.67
305
232.00
-73.00
220.17
190
271.67
81.67
237,67
Error
-84.83
47.67
Forecasting
Forecasting
Ft 1 (1 / N )
i t N 1
Di
t 1
(1 / N ) Dt Di Dt N
i t N
= Ft + (1 / N ) Dt Dt N
Forecasting
Forecasting
Period
Demand
MA(3)
MA(6)
10
12
14
10
16
12
18
14
11
10
20
16
13
11
22
18
15
12
24
20
17
Forecasting
Forecasting
Forecasting
That is,
New forecast = (Current observation of demand) + (1 )(Last forecast). In symbols,
Forecasting
Forecasting
Notice that since
Forecasting
(1
)
Dt i 1
i 0
= ai Dt i 1
i 0
where ai (1 ) i
and
ai
i 0
(1
)
= 1
i 0
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Quarter
Failures
Forecast
200
200
250
200
175
205
186
202
225
201
285
203
305
211
190
220
(by assumption)
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Differences:
The exponential smoothing forecast is a
weighted average of all past data points (as
long as the smoothing constant is strictly less
than 1).
The moving-average forecast is a weighted
average of only the last N periods of data.
Forecasting
Forecasting
Both exponential smoothing and movingaverage forecasts will lag behind a trend if
one exists.
Forecasting
Forecasting
Regression Analysis:
Let (x1, y1), (x2, y2), . , (xn, yn) be n paired
data points for the two variables X and Y.
Assume that yi, is the observed value of Y
when xi is the observed value of X. Refer to
Y as the dependent variable and X as the
independent variable.
Forecasting
Forecasting
Forecasting
S xy
S xx
a = D - b(n 1) / 2
Forecasting
Where,
n(n 1)
S xy n iDi
Di
2
i 1
i 1
and
n
n (n 1)(2n 1) n (n 1)
S xx
6
4
2
Forecasting
Forecasting
Forecasting
Then
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Period
Actual
Forecast
|Error|
200
200,00
10.00
200.00
0.00
250
209.00
9.90
218.90
31.10
175
222.01
10.21
232.22
57.22
186
226.50
9.64
236.14
50.14
225
231.12
9,14
240.26
15.26
285
238.74
8.98
247.72
37.28
305
251.45
9.36
260.81
44.19
190
265.23
9.80
275.02
85.02
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
The multiplier ct, represents the average
amount that the demand in the tth period of the
season is above or below the overall average.
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Week 1
Week 2
Week 3
Week 4
Monday
16.2
17.3
14.6
16.1
Tuesday
12.2
11.5
13.1
11.8
Wednesday
14.2
15.0
13.0
12.9
Thursday
17.3
17.6
16.9
16.6
Friday
22.5
23.5
21.9
24.3
Forecasting
Forecasting
Forecasting
Forecasting
Week l
Week 2
Week 3
Week 4
Monday
0.99
1.05
0.89
0.98
Tuesday
0.74
0.70
0.80
0.72
Wednesday
0.86
0.91
0.79
0.79
Thursday
1.05
1.07
1.03
1.01
Friday
1.37
1.43
1.33
1.48
Forecasting
Seasonal Factor
Monday
0.98
Tuesday
0.74
Wednesday
0.84
Thursday
1.04
Friday
1.40
Forecasting
Forecast
Monday
16.05
Tuesday
12.15
Wednesday
13.78
Thursday
17.10
Friday
23.05
Forecasting
Forecasting
Period
Demand
10
20
26
17
12
23
30
22
Forecasting
Forecasting
Forecasting
Period
Demand
Deseasonalized Demand
10
18.182
20
18.605
26
18.571
17
17.436
12
21.818
23
21.395
30
21.429
22
22.564
Forecasting
Notice that the deseasonalized demand
shows a clear trend. To forecast the
deseasonalized series one could use any of
the trend based methods discussed earlier.
Forecasting
Forecasting
Forecasting
Holt-Winters Method
We assume a model of the form
Dt = (+ Gt)ct + t.
Interpret as the base signal or intercept at
time t = 0 excluding seasonality, Gt as the trend
or slope component, ct , as the multiplicative
seasonal component in period t, and finally t
as the error term.
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
Forecasting
1
Ft
N
t 1
i t N
Forecasting
It follows that
t 1
E(Ft - Dt) = (1/N) E[ Di ] - E(Dt) =
i t N
(1/N)(N) - = 0.
This proves that when demand is stationary,
moving-average forecasts are unbiased.
Also,
Var(Ft - Dt) = Var(Ft) + Var(Dt)
Forecasting
t 1
= (1/N2)
Var ( D ) + Var(Dt)
i t N
= (1/N2)(N2) + 2
Forecasting
N 1
N
Forecasting
Forecasting
Since Ft is a linear combination of Dt-1, Dt2, ...., Dt-N, it follows that Ft is normally
distributed and independent of Dt.
Forecasting
Forecasting
E(Ft) = [ + (1 - ) + (1 - )2 + .] =
.
Notice that this means that E(et) = 0, so that
both exponential smoothing and moving
averages are unbiased forecasting methods
when the underlying demand process is a
constant plus a random term.
Forecasting
Var(Ft) = 22 + (1 - )222 +
= 22 1
2n
n 0
1
n0
2n
1
1 1
Forecasting
2 2
2
Var( Ft )
2
2
1 1
Since
Var(et) = 2 [/(2 - ) + 1]
= 2[2/(2 - )]
Forecasting
Or e =
2
2
Forecasting
= 2/(N + 1)