forecasting - rjerz.com · •describe three measures of forecast accuracy. ... choosing a...
TRANSCRIPT
1
Forecasting
Dr. Rick Jerz
© 2018 rjerz.com1
Learning Objectives
• Describe why forecasts are used and list the elements of a good forecast.
• Outline the steps in the forecasting process. • Describe at least three qualitative and three
quantitative forecasting techniques, and their advantages and disadvantages.
• Describe three measures of forecast accuracy.• Identify the major factors to consider when
choosing a forecasting technique.
© 2018 rjerz.com2
Making Decisions
• If we know exactly what will happen in the future, operational decisions would be easy.
• Problem – we don’t know the future.
© 2018 rjerz.com3
2
FORECAST
• A statement about the future
• Used to help managers• Plan the system• Operate the system
• Forecasting time horizons• Short-range• Medium-range• Long-range
????
© 2018 rjerz.com4
Forecasts
• Underlying basis of all business decisions!
© 2018 rjerz.com5
Department Uses of ForecastsAccounting Cost/profit estimatesFinance Cash flow and fundingHuman Resources Hiring/recruiting/trainingMarketing Pricing, promotion, strategyMIS IT/IS systems, servicesOperations Schedules, MRP, workloadsProduct/service design New products and services
© 2018 rjerz.com6
3
Forecasts Characteristics
• Assumes causal system and system stabilitypast ==> future
• Forecasts rarely perfect because of randomness
• Forecasts more accurate forgroups vs. individuals
• Forecast accuracy decreases as time horizon increases
• Statistics & math are often used
© 2018 rjerz.com7
Elements of a Good Forecast
Timely
AccurateReliable
Meaningful
WrittenEas
y to use
Timely
AccurateReliable
Meaningful
WrittenEas
y to use
© 2018 rjerz.com8
Seven Steps in Forecasting
1. Determine the use of the forecast2. Select the items to be forecasted3. Determine the time horizon of the forecast4. Select the forecasting model(s)5. Gather the data6. Make the forecast7. Validate and implement results
© 2018 rjerz.com9
4
Types of Forecasts
• Judgmental - uses subjective inputs• Time series - uses historical data assuming
the future will be like the past• Associative models - uses explanatory
variables to predict the future
© 2018 rjerz.com10
Judgmental Forecasts“Qualitative”
• Executive opinions• Sales force composite• Consumer surveys• Outside opinion• Opinions of managers and staff• Delphi method
© 2018 rjerz.com11
Time Series Forecasts“Quantitative”
• Trend - long-term movement in data• Seasonality - short-term regular variations in
data• Cycles – wavelike variations, usually more
than one year’s duration• Irregular variations - caused by unusual
circumstances• Random variations - caused by chance
© 2018 rjerz.com12
5
Forecast Pattern Examples
© 2018 rjerz.com13
Most CommonQuantitative Methods
• Naïve forecasts• Moving average• Weighted moving average• Exponential smoothing• Trend analysis
TimeTime--Series Series ModelsModels
TimeTime--Series Series ModelsModels
© 2018 rjerz.com14
Naïve Forecast
)(1 trendAF tt += -
• The forecast for any period equals the previous period’s actual value.
• Can be adjusted with trend
© 2018 rjerz.com15
6
Naïve Forecasts
• Advantages• Simple to use• Virtually no cost• Data analysis is nonexistent• Easily understandable• Can include trend and seasonality considerations
• Disadvantages• Cannot provide high accuracy• Can be a standard for accuracy
© 2018 rjerz.com16
Moving Average
nAAAF ttnt
t12... --- +++
=
• A technique that averages a number of recent actual values, updated as new values become available.
Moving average =Moving average = ∑∑ demand in previous n periodsdemand in previous n periodsnnMoving average =Moving average = ∑∑ demand in previous n periodsdemand in previous n periodsnn
∑∑ demand in previous n periodsdemand in previous n periodsnn
© 2018 rjerz.com17
Simple Moving Average Characteristics
• Must choose the number of periods, n• Only use the most recent “n” periods• Gets rid of old data• Premise: newer data is more indicative of
what the future will be• The bigger the “n”, the less “sensitive” the
forecast• The smaller the “n”, the more “reactive” the
forecast• Does not forecast trend well
© 2018 rjerz.com18
7
Moving Average: Periods
35373941434547
1 2 3 4 5 6 7 8 9 10 11 12
Actual
MA3
MA5
© 2018 rjerz.com19
Weighted Moving Average
å---- +++
=w
ttnntnt
AwAwAwF 1121...
• More recent values in a series are given more weight in computing the forecast.
WeightedWeightedmoving averagemoving average ==
∑∑ ((weight for period nweight for period n))x x ((demand in period ndemand in period n))
∑∑ weightsweightsWeightedWeighted
moving averagemoving average ==WeightedWeightedmoving averagemoving average ==
∑∑ ((weight for period nweight for period n))x x ((demand in period ndemand in period n))
∑∑ weightsweights
∑∑ ((weight for period nweight for period n))x x ((demand in period ndemand in period n))
∑∑ weightsweights
© 2018 rjerz.com20
Weighted Moving Average Characteristics
• Weights are usually a percent• Sum of the weights equals 1, or 100%• More complex than simple moving average• More responsive to most recent events• Weights based on experience and intuition• Better (than simple moving average) at
forecasting trend
© 2018 rjerz.com21
8
Exponential Smoothing
)( 111 --- -+= tttt FAFF a
• Premise -- The most recent observations might have the highest predictive value. Therefore, we should give more weight to the more recent time periods when forecasting.
New forecast =New forecast = last periodlast period’’s forecasts forecast+ + aa ((last periodlast period’’s actual demand s actual demand
–– last periodlast period’’s forecasts forecast))
© 2018 rjerz.com22
Exponential Smoothing Characteristics
• A form of weighted moving average• Requires a smoothing constant (alpha, α)
that ranges from 0 to 1 • The larger alpha, the more reactive the
forecast model.• Good at forecasting trend• Involves little record keeping of past data
© 2018 rjerz.com23
Trend Analysis
• Linear trends - Fitting a trend line to historical data points to project into the medium-to-long-range
• Linear trends can be found using the least squares technique
y y = = a a + + bxbx^̂y y = = a a + + bxbx^̂
where ywhere y = computed value of the variable to = computed value of the variable to be predicted (dependent variable)be predicted (dependent variable)
aa = y= y--axis interceptaxis interceptbb = slope of the regression line= slope of the regression linexx = the independent variable= the independent variable
^̂where ywhere y = computed value of the variable to = computed value of the variable to be predicted (dependent variable)be predicted (dependent variable)
aa = y= y--axis interceptaxis interceptbb = slope of the regression line= slope of the regression linexx = the independent variable= the independent variable
^̂
© 2018 rjerz.com24
9
Least Squares Method
Time periodTime period
Valu
es o
f Dep
ende
nt V
aria
ble
Time periodTime period
Valu
es o
f Dep
ende
nt V
aria
ble
DeviationDeviation11
DeviationDeviation55
DeviationDeviation77
DeviationDeviation22
DeviationDeviation66
DeviationDeviation44
DeviationDeviation33
DeviationDeviation11DeviationDeviation11
DeviationDeviation55DeviationDeviation55
DeviationDeviation77DeviationDeviation77
DeviationDeviation22DeviationDeviation22
DeviationDeviation66DeviationDeviation66
DeviationDeviation44DeviationDeviation44
DeviationDeviation33DeviationDeviation33
Actual observation Actual observation (y value)(y value)
Actual observation Actual observation (y value)(y value)
Trend line, y = a + bxTrend line, y = a + bx^̂Trend line, y = a + bxTrend line, y = a + bx^̂Trend line, y = a + bxTrend line, y = a + bx^̂
© 2018 rjerz.com25
Equations to Calculatethe Regression Variables
b =b = SSxy xy -- nxynxySSxx22 -- nxnx22
b =b = SSxy xy -- nxynxySSxx22 -- nxnx22SSxy xy -- nxynxySSxx22 -- nxnx22
y y = = a a + + bxbx^̂y y = = a a + + bxbx^̂
a = y a = y -- bxbxa = y a = y -- bxbx© 2018 rjerz.com26
Least Squares Example
b b = = = 10.5= = = 10.544SSxy xy -- nxynxySSxx22 -- nxnx22
3,063 3,063 -- (7)(4)(98.86)(7)(4)(98.86)140 140 -- (7)(4(7)(422))b b = = = 10.5= = = 10.544
SSxy xy -- nxynxySSxx22 -- nxnx22b b = = = 10.5= = = 10.544SSxy xy -- nxynxySSxx22 -- nxnx22
3,063 3,063 -- (7)(4)(98.86)(7)(4)(98.86)140 140 -- (7)(4(7)(422))
3,063 3,063 -- (7)(4)(98.86)(7)(4)(98.86)140 140 -- (7)(4(7)(422))
aa = = yy -- bxbx = 98.86 = 98.86 -- 10.54(4) = 56.7010.54(4) = 56.70aa = = yy -- bxbx = 98.86 = 98.86 -- 10.54(4) = 56.7010.54(4) = 56.70
TimeTime Electrical Power Electrical Power YearYear Period (x)Period (x) DemandDemand xx22 xyxy
19991999 11 7474 11 747420002000 22 7979 44 15815820012001 33 8080 99 24024020022002 44 9090 1616 36036020032003 55 105105 2525 52552520042004 66 142142 3636 85285220052005 77 122122 4949 854854
SSxx = 28= 28 SSyy = 692= 692 SSxx22 = 140= 140 SSxyxy = 3,063= 3,063xx = 4= 4 yy = 98.86= 98.86
TimeTime Electrical Power Electrical Power YearYear Period (x)Period (x) DemandDemand xx22 xyxy
19991999 11 7474 11 747420002000 22 7979 44 15815820012001 33 8080 99 24024020022002 44 9090 1616 36036020032003 55 105105 2525 52552520042004 66 142142 3636 85285220052005 77 122122 4949 854854
SSxx = 28= 28 SSyy = 692= 692 SSxx22 = 140= 140 SSxyxy = 3,063= 3,063xx = 4= 4 yy = 98.86= 98.86
TimeTime Electrical Power Electrical Power YearYear Period (x)Period (x) DemandDemand xx22 xyxy
19991999 11 7474 11 747420002000 22 7979 44 15815820012001 33 8080 99 24024020022002 44 9090 1616 36036020032003 55 105105 2525 52552520042004 66 142142 3636 85285220052005 77 122122 4949 854854
SSxx = 28= 28 SSyy = 692= 692 SSxx22 = 140= 140 SSxyxy = 3,063= 3,063xx = 4= 4 yy = 98.86= 98.86
© 2018 rjerz.com27
10
Least Squares Example
| | | | | | | | |19991999 20002000 20012001 20022002 20032003 20042004 20052005 20062006 20072007
160 160 –150 150 –140 140 –130 130 –120 120 –110 110 –100 100 –
90 90 –80 80 –70 70 –60 60 –50 50 –
YearYear
Pow
er d
eman
dPo
wer
dem
and
| | | | | | | | |19991999 20002000 20012001 20022002 20032003 20042004 20052005 20062006 20072007
160 160 –150 150 –140 140 –130 130 –120 120 –110 110 –100 100 –
90 90 –80 80 –70 70 –60 60 –50 50 –
| | | | | | | | |19991999 20002000 20012001 20022002 20032003 20042004 20052005 20062006 20072007
160 160 –150 150 –140 140 –130 130 –120 120 –110 110 –100 100 –
90 90 –80 80 –70 70 –60 60 –50 50 –
YearYear
Pow
er d
eman
dPo
wer
dem
and
Trend line,Trend line,y y = 56.70 + 10.54x= 56.70 + 10.54x^̂Trend line,Trend line,y y = 56.70 + 10.54x= 56.70 + 10.54x^̂Trend line,Trend line,y y = 56.70 + 10.54x= 56.70 + 10.54x^̂
© 2018 rjerz.com28
Trend Line Characteristics
• Great, if data has trend• Variations around the line are assumed to be
random• If trend is not linear, cannot use trend line• Trend could be curves, but math becomes
more difficult• Plot the data to insure a linear relationship• Be careful forecasting far beyond the
database
© 2018 rjerz.com29
More Advanced Techniques
• Associative forecasting• Predictor variables not “time”
• Multiple linear regression• More than one predictor variable
© 2018 rjerz.com30
11
Forecast Accuracy
• We generally do this by selecting the model that gives us the lowest forecast error
• Error - difference between actual value and predicted value
• Mean absolute deviation (MAD)• Average absolute error
• Mean squared error (MSE)• Average of squared error
• Mean absolute percentage error (MAPE)• Average of the percent errors
• Correlation coefficient – for trend line
© 2018 rjerz.com31
Mean Error
nForecastActual
ErrorMean å -=
|)(|
• Difference between actual and forecast• Absolute value used
© 2018 rjerz.com32
Mean Squared Error
• The error is squared• “Variance”
• Note: Some authors divide by n-1
© 2018 rjerz.com33
12
Mean AbsolutePercentage Error
• The absolute error is divided by the actual to calculate the percent errors
• These errors are averaged
© 2018 rjerz.com34
Correlation and Linear Regression
• Are two sets of data related?• Correlation analysis
• Is the relationship “linear”?• Linear regression analysis
• How strong is the relationship?• Correlation• Coefficient of Correlations (denoted by r).• Values range from -1 to +1• Values close to 0.0 indicate a weak correlation• Negative values indicate an inverse relationship
© 2018 rjerz.com35
Correlation
© 2018 rjerz.com36
13
Correlation
© 2018 rjerz.com37
Correlation Coefficient Equation to Calculate “r”
© 2018 rjerz.com38
Choosing a Forecasting Method
• Cost• Accuracy• Timely• Understandable• Serves purpose• In writing
© 2018 rjerz.com39
14
Operations Strategy
• Forecasts are the basis for many decisions• Work to improve short-term forecasts• Accurate short-term forecasts improve• Profits• Lower inventory levels• Reduce inventory shortages• Improve customer service levels• Enhance forecasting credibility
• Consider accountability of person making the forecast
© 2018 rjerz.com40