2. seasonal trends
peak periodsNumber of variables
7-*
Patterns in ForecastsTrend
A gradual long-term up or down movement of demand
Demand
Time
Upward Trend
7-*
Patterns in ForecastsCycle
An up and down repetitive movement in demand
Demand
Time
Cyclical Movement
7-*
Quantitative TechniquesTwo widely used techniques
Time series analysis
Linear regression analysisTime series analysis studies the
numerical values a variable takes over a period of timeLinear
regression analysis expresses the forecast variable as a
mathematical function of other variables
7-*
Time Series AnalysisLatest Period MethodMoving
AveragesExample ProblemWeighted Moving
3. AveragesExponential SmoothingExample Problem
7-*
Latest Period MethodSimplest method of forecastingUse
demand for current period to predict demand in the next
periode.g., 100 units this week, forecast 100 units next weekIf
demand turned out to be only 90 units then the following weeks
forecast will be 90
7-*
Moving AveragesUses several values from the recent past to
develop a forecastTends to dampen or smooth out the random
increases and decreases of a latest period forecastGood for
stable demand with no pronounced behavioral patterns
7-*
Moving AveragesMoving averages are computed for specific
periods
Three months
Five months
The longer the moving average the smoother the forecastMoving
average formula
7-*
Moving Averages - NASDAQ
4. 7-*
Weighted MAAllows certain demands to be more or less
important than a regular MAPlaces relative weights on each of
the period demandsWeighted MA is computed as such
7-*
Weighted MAAny desired weights can be assigned, but
SWi=1Weighting recent demands higher allows the WMA to
respond more quickly to demand changesThe simple MA is a
special case of the WMA with all weights equal, Wi=1/nThe
entire demand history is carried forward with each new
computationHowever, the equation can become burdensome
7-*
Exponential SmoothingBased on the idea that a new average can
be computed from an old average and the most recent observed
demande.g., old average = 20, new demand = 24, then the new
average will lie between 20 and 24Formally,
7-*
Exponential SmoothingNote: a must lie between 0.0 and
1.0Larger values of a allow the forecast to be more responsive
to recent demandSmaller values of a allow the forecast to
respond more slowly and weights older data more0.1 < a < 0.3
is usually recommended
5. 7-*
Exponential SmoothingThe exponential smoothing form
Rearranged, this form is as such
This form indicates the new forecast is the old forecast plus a
proportion of the error between the observed demand and the
old forecast
7-*
Why Exponential Smoothing?Continue with expansion of last
expressionAs t>>0, we see (1-a)t appear and <<1The demand
weights decrease exponentiallyAll weights still add up to
1Exponential smoothing is also a special form of the weighted
MA, with the weights decreasing exponentially over time
7-*
Forecasting with SeasonalityCalculate the average demand per
season
e.g.: average quarterly demandCalculate a seasonal index for
each season of each year:
Divide the actual demand of each season by the average demand
per season for that yearAverage the indexes by season
e.g.: take the average of all Spring indexes, then of all Summer
indexes, ...
7-*
Forecasting with SeasonalityForecast demand for the next year
6. & divide by the number of seasons
Use regular forecasting method & divide by four for average
quarterly demandMultiply next year’s average seasonal demand
by each average seasonal index
Result is a forecast of demand for each season of next year
7-*
Forecast ErrorError
Cumulative Sum of Forecast Error
Mean Square Error
7-*
Forecast ErrorMean Absolute Error
Mean Absolute Percentage Error
7-*
CFEReferred to as the bias of the forecastIdeally, the bias of a
forecast would be zeroPositive errors would balance with the
negative errorsHowever, sometimes forecasts are always low or
always high (underestimate/overestimate)
7-*
MSE and MADMeasurements of the variance in the
7. forecastBoth are widely used in forecastingEase of use and
understandingMSE tends to be used more and may be more
familiarLink to variance and SD in statistics
7-*
MAPENormalizes the error calculations by computing percent
errorAllows comparison of forecasts errors for different time
series dataMAPE gives forecasters an accurate method of
comparing errorsMagnitude of data set is negated
MA
n
=
D
i
n
å
i
=
1
to
n
,
n
=#
of periods in MA,
D
i
=
13. rights reserved.
Chapter 3
Probability and Statistics
A Foundation for Becoming a More Effective and Efficient
Problem Solver
3-*
Normal DistributionCommon probability distributione.g.,
height, weight, age, sum of two dice rolled 1,000 times, etc.
3-*
Normal Distribution
Sheet: Chart1
Sheet: Sheet1
Sheet: Sheet2
Sheet: Sheet3
Sheet: Sheet4
Sheet: Sheet5
Sheet: Sheet6
Sheet: Sheet7
Sheet: Sheet8
Sheet: Sheet9
Sheet: Sheet10
Sheet: Sheet11
Sheet: Sheet12
Sheet: Sheet13
Sheet: Sheet14
Sheet: Sheet15
Sheet: Sheet16
23. 3-*
Mean and Standard DeviationMost common statistics usedMean
or expected value
E(x) = SxiP(xi)
Standard deviation
s(x) = [S [xi - E(x)]2P(xi)]0.5
s(x) = [S [xi - m]2/n-1]0.5
3-*
Z-ScoresStandard Z-scoreMeasures the number of standard
deviations away from the meanCalculated as such:
Look up Z value in table to find probability
Normal Distribution
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0
1
2
3
4