Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Learning Objectives
Describe what forecasting is
Explain time series & its components
Smooth a data series
Moving average
Exponential smoothing
Simple
Auto-
What Is Forecasting?
Process of predicting a
future event
Underlying basis of
all business decisions
Production
Inventory
Personnel
Facilities
Forecasting Approaches
Qualitative Methods
Used when situation is
vague & little data exist
New products
New technology
Involve intuition,
experience
e.g., forecasting sales on
Internet
Quantitative Methods
Forecasting Approaches
Qualitative Methods
Used when situation is
vague & little data exist
New products
New technology
Quantitative Methods
Used when situation is
stable & historical data
exist
Existing products
Current technology
Involve intuition,
Involve mathematical
experience
techniques
e.g., forecasting sales on
e.g., forecasting sales of
Internet
color televisions
Quantitative Forecasting
Quantitative Forecasting
Methods
Quantitative Forecasting
Methods
Quantitative
Forecasting
Quantitative Forecasting
Methods
Quantitative
Forecasting
Time Series
Models
Quantitative Forecasting
Methods
Quantitative
Forecasting
Time Series
Models
Causal
Models
Quantitative Forecasting
Methods
Quantitative
Forecasting
Causal
Models
Time Series
Models
Moving
Average
Exponential
Smoothing
Trend
Models
Quantitative Forecasting
Methods
Quantitative
Forecasting
Causal
Models
Time Series
Models
Moving
Average
Exponential
Smoothing
Trend
Models
Regression
Quantitative Forecasting
Methods
Quantitative
Forecasting
Causal
Models
Time Series
Models
Moving
Average
Exponential
Smoothing
Trend
Models
Regression
Example
Year: 1995199619971998 1999
Sales: 78.763.589.7 93.292.1
Cyclical
Seasonal
Cyclical
Cyclical
Seasonal
Irregular
Trend Component
Persistent, overall upward or downward
pattern
Due to population, technology etc.
Several years duration
Response
Trend Component
t ren
d
r
a
w
Up
Time
Cyclical Component
Repeating up & down movements
Due to interactions of factors influencing
economy
Usually 2-10 years duration
Cycle
Response
Cyclical Component
Cycle
Time
Seasonal Component
Regular pattern of up & down fluctuations
Due to weather, customs etc.
Occurs within one year
Summer
Response
1984-1994 T/Maker Co.
Mo., Qtr.
Seasonal Component
Winter
Irregular Component
Erratic, unsystematic, residual fluctuations
Due to random variation or unforeseen
events
Union strike
War
Random or Irregular
Component
Erratic, Nonsystematic, Random,
Residual Fluctuations
No
Trend?
No
Trend?
Yes
Trend
Models
No
Trend?
Exponential
Smoothing
Yes
Trend
Models
NN oo
TT rr ee nn dd ??
YY ee ss
TT rr ee nn dd
MM oo dd ee ll ss
EE xx pp oo nn ee nn tt ii aa ll
SS mm oo oo tt hh ii nn gg
QQ uu aa dd rr aa tt ii cc
EE xx pp oo nn ee nn tt ii aa ll
AA uu tt oo -RR ee gg rr ee ss ss ii vv ee
No
trend? Yes
Smoothing
Forecasting
moving averages
linear
exponential smoothing
quadratic
exponential
autoregressive
Decomposition
cyclical analysis
10
Number of Passengers
0
09/83
07/86
05/89
Month/Year
Data collected by Coop Student (10/6/95)
03/92
01/95
NN oo
TT rr ee nn dd ??
YY ee ss
TT rr ee nn dd
MM oo dd ee ll ss
EE xx pp oo nn ee nn tt ii aa ll
SS mm oo oo tt hh ii nn gg
QQ uu aa dd rr aa tt ii cc
EE xx pp oo nn ee nn tt ii aa ll
AA uu tt oo -RR ee gg rr ee ss ss ii vv ee
8
6
4
2
0
93
94
95
96
Year
97
98
Moving Average
[An Example]
You work for Firestone Tire. You
want to smooth random
fluctuations using a 3-period
moving average.
1995
20,000
1996
24,000
1997
22,000
1998
26,000
1999
25,000
Moving Average
[Solution]
Year
1995
1996
1997
1998
1999
Moving Average
Year
Response Moving
Ave
1994
NA
1995
1996
1997
3.67
1998
1999
NA
Sales
8
6
4
2
0
94 95 96 97 98 99
Exponential Smoothing
Method
NN oo
TT rr ee nn dd ??
YY ee ss
TT rr ee nn dd
MM oo dd ee ll ss
EE xx pp oo nn ee nn tt ii aa ll
SS mm oo oo tt hh ii nn gg
QQ uu aa dd rr aa tt ii cc
EE xx pp oo nn ee nn tt ii aa ll
AA uu tt oo -RR ee gg rr ee ss ss ii vv ee
Exponential Smoothing
Method
Form of weighted moving average
Weights decline exponentially
Most recent data weighted most
Exponential Smoothing
[An Example]
Youre organizing a Kwanza meeting. You
want to forecast attendance for 1998 using
exponential smoothing
( = .20). Past attendance (00) is:
1995
4
1996
6
1997
5
1998
3
1999
7
Exponential Smoothing
Ei = WYi + (1 - W)Ei-1
Smoothed Value, Eii
(W = .2)
Forecast
^
Yii ++ 11
Time
Yii
1995
4.0
NA
1996
4.0
1997
4.4
1998
4.5
1999
4.2
2000
NA
NA
4.8
Exponential Smoothing
[Graph]
Attendance
Actual
8
6
4
2
0
93
96
97
Year
98
99
Forecast Effect of
Smoothing Coefficient (W)
^
Yi+1 = WYi + W(1-W)Yi-1 + W(1-W)2Yi-2 +...
Weight
2 Periods 3 Periods
W is... Prior Period
Ago
Ago
22
W(1-W)
W(1-W)
0.10
10%
9%
8.1%
0.90
90%
9%
0.9%
Linear Time-Series
Forecasting Model
NN oo
TT rr ee nn dd ??
YY ee ss
TT rr ee nn dd
MM oo dd ee ll ss
EE xx pp oo nn ee nn tt ii aa ll
SS mm oo oo tt hh ii nn gg
QQ uu aa dd rr aa tt ii cc
EE xx pp oo nn ee nn tt ii aa ll
AA uu tt oo -RR ee gg rr ee ss ss ii vv ee
Linear Time-Series
Forecasting Model
Used for forecasting trend
Relationship between response variable Y &
time X is a linear function
Coded X values used often
Year X:
Coded year:
Sales Y:
1999
4
92.1
Yi b0 b1X 1i
b1 > 0
b1 < 0
TT ii mm ee ,, XX 11
1
1
2
2
4
Coded Year
0
1
2
3
4
5
Sales (Units)
1
1
2
2
4
?
Yi b0 b1 X i 2.143 .743 X i
94
95
96
97
98
99
Excel Output
Coefficients
Intercept
2.14285714
X Variable 1 0.74285714
8
7
Projected to
year 2000
2
1
0
1993
1994
1995
1996
1997
1998
1999
2000
(X 1000)
20
19
18
Number of Surgeries
17
16
01/93
01/94
01/95
Month/Year
Source:
01/96
01/97
(X 100)
193
191
189
Number of Surgeries
187
185
183
01/93
01/94
01/95
Month/Year
Source: General Hospital, Metropolis
01/96
01/97
Seasonality Plot
Revised Surgery Data
(Seasonal Decomposition)
100.5
100.3
Monthly Index
100.1
99.9
99.7
Jan
Feb
Mar
Apr
May
Jun
Jul
Month
Source:
Aug
Sep
Oct
Nov
Dec
Trend Analysis
Revised Surgery Data
(Trend Analysis)
(X 1000)
19.5
19.2
18.9
Number of Surgeries
18.6
18.3
18
12/92
10/93
8/94
6/95
Month/Year
Source: General Hospital, Metropolis
9/96
2/97
12/97
Quadratic Time-Series
Forecasting Model
NN oo
TT rr ee nn dd ??
YY ee ss
TT rr ee nn dd
MM oo dd ee ll ss
EE xx pp oo nn ee nn tt ii aa ll
SS mm oo oo tt hh ii nn gg
QQ uu aa dd rr aa tt ii cc
EE xx pp oo nn ee nn tt ii aa ll
AA uu tt oo -RR ee gg rr ee ss ss ii vv ee
Quadratic Time-Series
Forecasting Model
Used for forecasting trend
Relationship between response variable Y &
time X is a quadratic function
Coded years used
Quadratic Time-Series
Forecasting Model
Used for forecasting trend
Relationship between response variable Y &
time X is a quadratic function
Coded years used
Quadratic model
2
Yi b0 b1X 1i b11X 1i
Quadratic Time-Series
Model Relationships
Y
b11 > 0
Y e a r, X
Y
Y e a r, X
Y e a r, X
b11 < 0
b11 > 0
b11 < 0
Y e a r, X
95
96
97
98
99
Yi b0 b1 X i b2 X i
Coefficients
Intercept
2.85714286
X Variable 1 -0.3285714
X Variable 2 0.21428571
Excel Output
2
Exponential Time-Series
Model
NN oo
TT rr ee nn dd ??
YY ee ss
TT rr ee nn dd
MM oo dd ee ll ss
EE xx pp oo nn ee nn tt ii aa ll
SS mm oo oo tt hh ii nn gg
QQ uu aa dd rr aa tt ii cc
EE xx pp oo nn ee nn tt ii aa ll
AA uu tt oo -RR ee gg rr ee ss ss ii vv ee
Exponential Time-Series
Forecasting Model
Used for forecasting trend
Relationship is an exponential function
Series increases (decreases) at increasing
(decreasing) rate
Exponential Time-Series
Forecasting Model
Used for forecasting trend
Relationship is an exponential function
Series increases (decreases) at increasing
(decreasing) rate
Exponential Time-Series
Model Relationships
b1 > 1
0 < b1 < 1
Y e a r, X
Exponential Weight
[Example Graph]
Sales
8
Data
6
4
Smoothed
2
0
94
95
96
97
98
99 Year
Yi b0b1
94
95
Coefficients
Intercept 0.33583795
X Variable 10.08068544
96
97
98
antilog(.33583795) = 2.17
antilog(.08068544) = 1.2
99
Xi
Yi ( 2.17 )( 1.2 )
Autoregressive Modeling
NN oo
TT rr ee nn dd ??
YY ee ss
TT rr ee nn dd
MM oo dd ee ll ss
EE xx pp oo nn ee nn tt ii aa ll
SS mm oo oo tt hh ii nn gg
QQ uu aa dd rr aa tt ii cc
EE xx pp oo nn ee nn tt ii aa ll
AA uu tt oo -RR ee gg rr ee ss ss ii vv ee
Autoregressive Modeling
Used for forecasting trend
Like regression model
Independent variables are lagged response
variables Yi-1, Yi-2, Yi-3 etc.
10
Number of Passengers
0
09/83
07/86
05/89
Month/Year
Data collected by Coop Student (10/6/95)
03/92
01/95
Auto-correlation Plot
Intra-Campus Bus
(Auto Correlation
PassengersFunction
Plot
0.5
-0.5
-1
0
10
15
Lag
20
25
Autoregressive Model
[An Example]
4
3
2
3
2
2
4
6
Autoregressive Model
[Example Solution]
Develop the 2nd order
table
Use Excel to run a
regression model
Excel Output
Coefficients
Intercept
3.5
X Variable 1
0.8125
X Variable 2
-0.9375
Year
92
93
94
95
96
97
98
99
Yi
4
3
2
3
2
2
4
6
Yi-1
--4
3
2
3
2
2
4
Yi-2
----4
3
2
3
2
2
Evaluating Forecasts
Quantitative
Forecasting Steps
Select several forecasting methods
Forecast the past
Evaluate forecasts
Select best method
Forecast the future
Monitor continuously forecast accuracy
Forecasting Guidelines
No pattern or direction in forecast error
ei = (Actual Yi - Forecast Yi)
Seen in plots of errors over time
Simplest model
Called principle of parsimony
Desired Pattern
EE rr rr oo rr
EE rr rr oo rr
00
00
TT ii mm ee (( YY ee aa rr ss ))
TT ii mm ee (( YY ee aa rr ss ))
Residual Analysis
e
Random errors
T
Cyclical effects not accounted for
T
Seasonal effects not accounted for
Principal of Parsimony
Summary
Described what forecasting is
Explained time series & its components
Smoothed a data series
Moving average
Exponential smoothing
Estimation
[Know mechanics of
StatGraphics Plus Win].
Diagnostic checking
Questions?
ANOVA
End of Chapter