section 7.2: exponential smoothing
DESCRIPTION
Section 7.2: Exponential Smoothing. Quantitative Decision Making 7 th ed By Lapin and Whisler. Simple Exponential Smoothing. Graphing Actual vs Forecast Values. Forecasting Errors. Two Parameter Smoothing. Simple Exponential Smoothing. Compute T 3. Compute b 3. - PowerPoint PPT PresentationTRANSCRIPT
Section 7.2: Exponential Smoothing
Quantitative Decision Making 7th edBy Lapin and Whisler
Simple Exponential Smoothing
Year t Yt Ft
2000 1 2014 -------2001 2 1648 20142002 3 1694 19042003 4 2115 18412004 5 2167 19232005 6 2410 19962006 7 2464 21202007 8 2145 22232008 9 1210 22002009 10 270 -------
tt1t F)1(YF
2.19042014*7.1648*3.
Graphing Actual vs Forecast ValuesSummit County Sales Activities
0
500
1000
1500
2000
2500
3000
Nu
mb
er
of
Sa
les
Time
Acutal Sales Forecasted Sales
Forecasting Errors
n
ttt FY
nMAD
1
1
n
t t
tt
YFY
nMAPE
1
100
2
1
)(1t
n
tt FY
nMSE
Two Parameter Smoothing
)bT)(1(YT 1t1ttt
1t1ttt b)1()TT(b
tt1t bTF
Simple Exponential Smoothing
Year t Yt Tt bt Ft
2000 1 2014 ------- -------- -------
2001 2 1648 2014 1648-2014
-366-------
2002 3 1694 2014-366
1648
2003 4 2115
Compute T3
1661)3662014(7.1694*3.
)bT(7.Y3.T 2233
Year t Yt Tt bt Ft
2000 1 2014 ------- -------- -------2001 2 1648 2014 -366 -------
2002 3 1694 1661 16482003 4 2115
Compute b3
363)366(*8.)20141661(2.
b)1()TT(b 2233
Year t Yt Tt bt Ft
2000 1 2014 ------- -------- -------
2001 2 1648 2014 -366 -------
2002 3 1694 1661 -363 1648
2003 4 2115
Seasonal Exponential Smoothing with Three Parameters Many time series have regular seasonal patterns to
be incorporated into forecasts. The three-parameter model incorporates a
seasonal smoothing constant (beta):
Tt = Yt /St –p) + (1 – )(Tt –1 + bt –1)
bt = (Tt – Tt –1) + (1 – )bt –1
St = Yt /Tt) + (1 – )St –p
Ft+1 = (Tt + bt) St –p+1
Forecasting withThree Parameters
Forecasting withThree Parameters The above works for p = 4 quarters or p =
12 months. The preceding slide needs 6 quarters to
generate the first (very bad) forecast. The process settles quickly, providing
good forecasts p periods into the future.