By:
Dr. Abdulfatah Salem7 Predict or estimate a
Is the practice of predicting something that will
happen or come to be A the future. It is not,
however, like or other
parton ese fens Poy eenot perched
Forecasting relies on solid evidence to predict
future events. It is used in a number of areas
today, most commonly in the process of
predicting the and the imminent
Gree Cs Meu Cont) Forecasting is the attempt to predict future
eC en outcomes based on past events
planning, and estimating future growth. Dee uraBecause many of the decisions be made under uncertainty
Forecasting lays a ground for reducing the risk in all
decision making
In all business applications, forecasting serves as a
starting point of the major decisions in all
organization functions
9-Jul-23 Dr. Abdulfatah Salem SSSSsS=e}New Laws and
Regulations
Competition Technology
Growth Trend
Political Change
9-Jul-23 Dr. Abdulfatah Salem | 4Darel f
Treg Lael 4
new product quality Pret ad
fesearch and i) improvement Perel
ROTI Tat lalg
= 9-Jul-23 Dr. Abdulfatah Salem [i 5) lB|Make forecast and implement results
99-Jul-23 E Dr. Abdulfatah Salem | gl| Facts about the forecasting |
{ A statement about the future value of a variable of interest such as demand. ]
{ Forecasting is used to make informed decisions.
{ People make and use forecasts all the time both in their jobs and in everyday lite
| Accurate forecasts lead to efficient operations
| Inaccurate forecasts lead to inefficient, high cost operations.
9-Jul-23 Dr. Abdulfatah SalemSZ ao
Forecasts rarely perfect because of randomness
25
LF
Forecast accuracy decreases as time horizon increases
a 9-Jul-23
Dr. Abdulfatah Salem 8*Consists of Objective inputs Consists of Subjective inputs
*Generated thru math. Modeling Generated by the forecaster
*Involve hard data «Includes soft information (human
(projection of historical data, factors, personal opinion, expertise )
analyzing objectives) used to predict what something in the
* used to predict future figures and future will be in terms of things other
quantities such as sizes and lengths. than set figures.
9-Jul-23 MM Dr: Abdulfatah SalemAveraging
Techniques
| 9-Jul-23_ | Dr. Abdulfatah SalemA method of forecasting using a composite
forecast prepared by a number of individual
executive.
The executives form their own opinions initially
from the data given, and revise their opinions
according to the others' opinions. Finally, the
individuals’ final opinions are combined.
Production manager 125 40% 50
marketing manager 160 35% 56
Financial manager 100 25% 25
9-Jul-23 Dr. Abdulfatah Salem 11Executive Opinion
Advantages
Disadvantages
1. This method is useful for a company which is
innovating new products.
1. Its opinion and experience oriented not based on
facts.
2, Very quick and simple process and itis easily
done through holding meetings periodically.
2. There can be biased opinions.
3. Collective decision is been taken considering the
opinions of all the top level executives.
4, Useful for small business because of the
simplicity of this method.
3. The responsibility of the decision is distributed
among all the executives and no single person will
be liable for any faulty forecast.
9-Jul-23
Dr. Abdulfatah SalemDelphi is based on the principle that forecasts (or decisions) from a structured group of individuals are
‘more accurate than those from unstructured groups. The technique can also be adapted for use in face-to-face
meetings, and is then called mini-Delphi or Estimate-Talk-Estimate (ETE). Delphi has been widely used for busine
forecasting and has certain advantages over another structured forecasting approach, prediction markets
| The experts answer questionnaires in two or more rounds. After each round, a facilitator or change agent provides an
anonymous summary of the experts’ forecasts from the previous round as well as the reasons they provided for their
judgments. Thus, experts are encouraged to revise their earlier answers in light of the replies of other members of their
panel, itis believed that during this process the range of the answers will decrease and the group will converge towards
the "correct" answer.
i Finally, the process is stopped after a predefined stop criterion (e.g. number of rounds, achievement of
consensus, stability of results) and the mean or median scores of the final rounds determine the results.
9-Jul-23 Dr: Abdulfatah Salem 13 |FORECASTING
9-Jul-23 Dr. Abdulfatah Salem | 14Time Series Forecast (Extrapolative)
Statistical techniques that Time ordered sequence of
make use of historical data observations taken at regular
collected over a long period of observations taken at regular
intervals.
Forecasts based on only one factor - time.
For example:
OGross Domestic Product Each Quarter
QWMonthly Profits
QO Annual Rainfall
QDaily Stock Market Index
9-Jul-23 Dr. Abdulfatah Salem | 16
time..FORECASTING
Patterns of the Time Series
Increasing
No Recognizable Pattern Linear Trend
Time, | Time,
} Seasonal Pattern plus
Linear Growth
2
Time,FORECASTING
Time Series fmeeenes
BEstEb ~— [Is S
Dr. Abdulfatah SalemFORECASTING
__Naiive Forecasts |
Estimating technique in which the last period's actual are used as this period's forecast,
| without adjusting them or attempting to establish causal factors. It is used only for comparison
with the forecasts generated by the better (sophisticated) techniques.
Example
Month
Sales actual
Forecast
JAN
390
FEB
420
MAR
190
APR
380
MAY
400
June
FeAu + (As# Ad)
June
9-Jul-23
Dr. Abdulfatah Salem | 19Simple Mean Forecasts
Ana = Ayn
Period
‘January,
February
Dr. Abdulfatah Salem | 20F, = MA,=
F, = forecast for time period t
MA, =n period moving average
A,.4 = actual value in period t-1
n_ number of periods ( data
points )
9-Jul-23 Dr. Abdulfatah Salem | otFORECASTING
‘he weekly sales figures (in millions of dollars)
presented in the following table are used by a
major department store to determine the need for
temporary sales personnel (n=3).
Determine the forecast value of week 26
5.7
Aegan
| 9-Jul-23_ | Dr. Abdulfatah Salem
Period @
a
la
|
la
cof
on /or}OJon}on| onl a jonionio
co] |] | 0 [e209
2] |] on} n] on} en1 > |en/o)
|S on|io Sfi
o|January
February
March
3 month MA:
(oct+nov+dec)/3 = 258.33
April
May
June 6 month MA:
July (jultaug+...tdec)/6 = 249.33
AUGUSE 12 month MA:
September (Jan+feb+...+dec)/12 = 205.33
October
November
December
9-Jul-23 Dr. Abdulfatah SalemHistorical values of the time series are assigned different weights
when performing the forecast.
In a weighted moving average, weights are assigned to the most
9-Jul-23 Dr. Abdulfatah Salem | 24The table contains the pas patelofichanee)
exchange rate between the Kuwaiti 20/8/2023) pO 7a
‘ . 21/4/2023 100.768
dinar and the Eqyptian pound from 22/4/2023 ADOnSe
20/4/2023 to 1/5/2023 Calculate the — | -szyazaoas a
four-days weighted moving 24/4/2023 100.566
average to forecast for the rate of 25/4/2023 101.256
change at 2/5/2023 using the 26/4/2023 100.668.
ights (0.62 0.21 0.13 0.04) 27/4/2023 99.898
28/4/2023 100.457
29/4/2023 100.893
e forecast value at 023 30/4/2023 101.187
= 101.213*0.62 + 100.785*0.21 31/4/2023 100.785
+ 101.187*0.13 + 100.893*0.04 erei2i23 a
= 101.107 2/5/2023 | 101.107
Lia3-period weighted moving average forecasting
Period Actual value
Month 1
Month 2
Month 3.
Month 4
Month 5
Month 6
Month 7
Month 8.
Month 9
Month 10
Month 11
Month 12
STARCO supply
Forecast
Dr. Abdulfatah Salem+The most recent observations might have the highest predictive value.
- Therefore, we should give more weight to the more recent time periods when forecasting.
-The method based on previous forecast plus a percentage of the forecast error
-Forecast = previous forecast + % Error
-Error = Actual value - Forecast value
is Fea)
Determination of OL is usually judgmental and subjective and often based on triakand -error
experimentation. The most commonly used values of Ol are between 0.10 and 0.50.
[__9-Jul-23__| Dr. Abdulfatah SalemFORECASTING
The shown table is the expected value and the actual value of the sales of the new
apartments this year, determine the forecast value of sales for the next year using exponential
smoothing method with a = 0.15
This year
Forecast Actual
$ 300000 $ 280000
Fpextyear = 300000 + 0.15 (280000 - 300000)
= 300000 + 0.15 (-20000)
= $§ 297000
9-Jul-23 Dr. Abdulfatah Salem | asGiven the weekly demand data, what are the exponential smoothing
forecasts for Month 10 using a = 0.10
1 #0 (Ary Fya) cee [ex
F, = 820+ 0.10 (820-820) = 820
F, = 820+ 0.10 (775-820) = 815,
F,= 815+ 0.10 (680-815) = 801.5
F, = 801.5 + 0.10 (655-801.5) - 787
F,= 787 +0.10 (750-787) = 783
F, = 783 + 0.10 (802-783) = 785
F,= 785 + 10 (798-785) = 786.3
Fy = 786.3 + 0.10 (689-786.3) = 776.6
Blelal|slafulaluln|e
F,p= 776.6 + 0.10 (775-7766) = 776.4
9-Jul-23 Dr. Abdulfatah Salem | 29[) Trend is a long term movement in a time series. It is the underlying direction(an upward
or downward tendency) and rate of change in a time series, when allowance has been
made for the other components. A simple way of detecting trend in seasonal data is to
take averages over a certain period and if these averages change with time we can say
that there is evidence of a trend in the series, indicating that the population mean is
time dependent.
F, = Forecast for period t
t= Specified number of time periods
a=Value of F, att=0
b = Slope of the line
9-Jul-23 Dr. Abdulfatah Salem | 30sFORECASTING
Calculating a and b
Time (t) | Value (y)
8
14
21
#
#
2)
paDZe)-ztzy
oes ;
att -(=bFORECASTING
weeks
addy. Et Dy
mrt? - (E92
= Ly -bEt
b= $0) 18012) es 12180
58 BS Tens
s12_- 6303)
= 3s
\ 43.5 + 6.3
\ )
| 9-Jul-23 | Dr. Abdulfatah Salem iiaThe shown are the electric Power consumption Year pElcctricalPower
for Alexandria city from 2012 to 2018. Sar _Demand (megawatt)
Required
Calculate the predicted demand for 2019 ?
First, plot the relation between the year and the consumptions.
Note that the relation is an ascending trend
Time Electrical Power
Year Period (¢) Demand (megawatt)
2012 74
2013 2 ”
2014 80
2015 90
2016 105
2017 142
2018 an
Sys@n
| 9-Jul-23 | Dr. Abdulfatah SalemnS (ty) -TtEy
nse. (507
70 + 10.54t
Trend line,
Y= 56.70 + 10.54t
2012 2013 2014 2015 2016 2017 2018 2018 2020
Year
| 9-Jul-23 | Dr. Abdulfatah Salem ise 34 as90,000,000,000
87,025,000,000
108,900,000,000
119,025,000,000
102,400,000,000
136,900,000,000
144,400,000,000
160,000,000,000
148, 225,000,000
184,900,000,000
1D | 0 [co |N Jo fur] for [ny | let
Demand for 2001 is y:
t=11 foryear2001
77333 + 14576*11 = 437669
9-Jul-23 MD. Abdulfatah SalemFORECASTING
Compu-Arab Computer assembly Co., It assemble computers from generic parts.
the company has had steady growth since it started. It purchase generic parts in volume at a
discount from a variety of sources whenever they see a good deal. Thus, they need a good
forecast of demand for their computers so that they will know how many parts to purchase and
stock. They have compiled demand data for the last 12 months as shown.
-patDity DID)
a
aye (50)
b=1.72 a=35.2
y=35.241.72t
for period 13.
8 y=35.2+1.72013)=57.56
| 9-Jul-23 | Dr. Abdulfatah Salem P3671912023 Dr. Abdulfatah SalemFORECASTING
Seasonality
Seasonality is a characteristic of a time series in which the data experiences regular
and predictable changes that recur every fixed interval of time. Any predictable change or
pattern in a time series that recurs or repeats over a one-time period can be said to be seasonal.
If you live in a climate with cold winters and warm summers, your home's heating costs probably
risen he winter and fall in the summer. You reasonably expect the seasonality of your heating costs to recur
every year.
| Seasonal
Seasoal atern
Linear GreTerms & Definitions
Seasonality
— Repetition at Fixed Intervals
‘Seasonal variations
— Regularly repeating movements in series values that can be tied to recurring events.
— Itis applied to annually, monthly, weekly, daily and other regularly recurring patterns
in data.
Seasonal relative (index)
— Percentage of average or trend
Centered moving average
— Amoving average positioned at the center of the data that were used to compute it.
7/9/2023 Dr. Abdulfatah Salem | 39FORECASTING
FORECASTING METHODS FOR SEASONAL SERIES ]
House Cleaning Services Company needs a quarterly forecast of the
number of customers expected next year. The business is seasonal, with a
peak in the third quarter and a trough in the first quarter.
Forecast customer demand for each quarter of year 5, based on an estimate
of total year 5 demand of 2,600 customers.
Quarter Year 1 Year 2 Year 3 Year 4
45, 70 100 100
335 370 585 725
520 590 830 1160
100 170 285 215
1000 1200 1800 2200
250 300 450 550
7/9/2023 Dr. Abdulfatah Salem | 40Multiplicative Seasonal Influences
A seasonal influence is multiplicative if the quarterly demand forecast of a
quarter
= Projected average period demand x Average seasonal index of that quarter.
Actual Seasonal Value
Average Value oi the Period
Seasonal Index =
Quarter Year 1 Year 2 Year 3 Year 4
1 45/250 = 0. 70/300 = 0. 100/450 =0.22 100/550 = 0.18
2 335/250 = 1.; 370/300 = 1. 585/450 = 1.30 725/550 = 1.32
3 520/250 = 2.1 590/300 = 1: 830/450 = 1.84 1160/550 = 2.11
4 100/250 = 0.: 170/300 = 0. 285/450 = 0.63 215/550 = 0.39
7/9/2023 Dr. Abdulfatah Salem | atSra
Bi Bes rite Sooke
Eee | Scull Forecast
Peeeenr| eC kcos een iol
demand for CES ela) Ta}
each period. [f) for each seasons of
period. Urol alelen
Dr. Abdulfatah Salem
7/9/2023FORECASTING
Quarterly demand for last four years is given in the table below.
use a 5-step process to forecast.
Quarter Year | Year 2
Fall 2530 2690
‘Winter 2300 2420
Spring 2000
Summer 1775
7/9/2023 Dr. Abdulfatah SalemFORECASTING
Step 1: Find average quarterly demand for each year.
Quarter Year 1 Year 2 Year 3 Year 4
Fall 2530 2690 2790 2860
Winter 2300 2420 2410 2600
Spring 1900 2000 2105 2175
Summer 1510 1775 1875 1945
Average 2060 2221 2295 2395
= (1510+ 1900+ | =(1775+2000+ | =(1875+2105+ = (1945 +2175 +
Formula 2300 + 2530)/4) 2420 + 2690)/4) 2410 + 2790)/4) 2600 + 2860)/4)
= 2060 = 2221 = 2295 = 2395
7/9/2023 Dr. Abdulfatah Salem 44FORECASTING
Step 2: Compute Seasonal Index (SI) for each quarter for each year.
Quarter
Fall
Winter
Spring
Summer
7/9/2023 Dr. Abdulfatah Salem | 45St. Pp 3: Calculate the average SI for each quarter.
Si, Si, Si, Si,
1.228 1.211 1.216 1.194
4117 1.090 1.050 1.086
0.922 0.900 0.917 0.908
0.733 0.799 0.817
7/9/2023 Dr. Abdulfatah Salem | 46Step 4: Calculate the average quarterly demand for next year.
First, the yearly demand has to be estimated or calculated for next year using
one of the forecasting techniques.
Year 1 Year 2 Year 3 Year 4 Year 5
2060 2221 2295 2395
a
Year1 Year? Year3 Year4 Year 5
7/9/2023 Dr. Abdulfatah SalemFORECASTING
Step 5: Forecast demand for the four quarters of next year.
Multiply the average demand by the SI for each quarter.
Fall 1.212 2599x1.212
Winter 1.086 2599x1.086
Spring 0.912 2599x0.912
Summer 0.790 2599x0.790
7/9/2023 Dr. Abdulfatah Salem | 48The manager of a parking lot has computed daily relatives for the number of cars per day in the
lot. The computations are repeated here (about three weeks are shown for illustration). A seven
period centered moving average is used because there are seven days (seasons) per week
Deny [rom ]ves [reer] rt] or [oon] wom frase] on [mor] rr [ont [oun un [ton] wos row] eri [one Joon [oe
Cars [|=] =| | |] =|] |||] o| ala] |r] |o]«|o
ron] wes frir| ei J 8 [ Son [ton one] wou] rr [ri [sot] in| won rome] wes] mar] re [ew | nf won
a|lel=|=|=|=[o|> =l«lel«|=|ol=[=|«|@
Centered MAT L [re[maroe] n [rafoelna manana] elnelra[e [ra] | |
Seasonal inde 1.36] +27] 0.51[0.77] 06] 1.03] 1.19 | 14 ]+.21]0.55]0.72]0.0] 1.06] +2 [1.33] I
Abdulfatah Salem = :The estimated relatives will be:
Monday: (0.77 + 0.72)i2
Tuesday: (0.
Wednesday: (1.03 + 1.06)2 = 1.045
Thursday: (1.19 + 1.2)/2 = 1.195
Friday: (1.36 + 1.4 + 1.33)/3 = 1.36
Saturday: (1.27 + 1.21)/2= 1.24
‘Sunday: (0.51 + 0.56)/2 = 0.535,
lote
The sum of the relatives must equal the number of periods (i.e., 7 in this example). If it is not, you
have to multiply by a correction factor. In this example the sum is 6.985, therefore you have to
multiply each factor by (7/6.985 = 1.00215)
746 Tue = 0.866 . Thu = 1.196
365 Sat=1.245
9-Jul-23 Dr. Abdulfatah Salem | 50FORECASTING
Days
cars
Weekly avg.
Next week
avg.
Next week forecast
Days
Relatives
Next week avg.
Seasonal value
9-Jul-23
Dr. Abdulfatah SalemFORECASTING
Associative Forecasting
Associative models assume that there is a causational relationship
between the variable of interest and other variables called
predictors
Predictor variables - used to predict values of variable
Price of beef and price of chicken of interest
Crop yields and soil condition Regression - technique for fitting a line to a set of points
Crop yields & timing of water Regression Methods ~ Regression (or causal ) methods
Profits and sales that attempt to develop a mathematical relationship
Price of products and energy cost between the item being forecast and factors that cause it
to behave the way it does.
7/9/2023 Dr. Abdulfatah Salem | 52FORECASTING
Simple Linear Regression
The simplest and most widely used form of regression involves a linear relationship
between two variables.
The objective in linear regression is to obtain an equation of a straight line that
minimizes the sum of squared vertical deviations of data points from the line (i.e.
the least squares criterion)
= nEGy)- Exdy
This least squares line has the form: nsx? - (Ext
y=a+bx
Where:
Y= predicted (dependent or response) variable -
predictor (dependent or explanatory) variable = Ex-th
tercept
lope (change rate)
umber of paired observations _/
7/9/2023 Dr. Abdulfatah Salem | 53Conditions required for predictor to be valid
|. The relationship between movements of an indicator and movements of the
variable should have a logical explanation.
. Movements of the indicator must precede movements of the dependent variable
by enough time so that the forecast isn’t outdated before it can be acted upon.
. A fairly high correlation (r ) should exist between the two variables.
T = 220y)- Exedy
er? 27) (027- y?)
Correlation factor r is'a measure of the strength and direction of relationship between two variables
Correlation r ¢ dicate weak linear relationship between two variables
“Correlation r le strength linear relationship between two variables (directly change)
“Correlation r cae strength linear relationship between lwo variables (inversely change)
7/9/2023 [ini 54 aaaFORECASTING
Simple Linear Regression
— Variations around the line are — Always plot the data to verify
random linearity
— Deviations around the line — Check for data being time-
normally distributed dependent
— Predictions are being made — Small correlation may imply
only within the range of that other variables are
observed values important
7/9/2023 Dr. Abdulfatah Salem | 55Carrefour has a chain of
10 Stores in Egypt. Sales figures and
profiles for the stores are giving in
the following table. Obtain a
regression for the data, and predict
profit for a store assuming sales of
23 million.
7/9/2023 Dr. Abdulfatah Salem | 56441
324
361
484
256
170 289 100
375 625 226
20 13 260 400 169
ExEI9O Sy=1115 = Sxy=2184 YxX=3694 Sy=1297.25
pr -—B zy) dad
(es? -") xs
I= 093
7/9/2023 Dr. Abdulfatah Salempb = 10(2184) — (190)4111.5
10(3694) — (190)2
_ 111.5—0.78 (190) _
10
-3.67 + 0.78X
when x = 23 million
-3.67 + 0.78(-23)
14.27 million
PgThe following table shows the profits (in EGP)
resulting from the weekly sales (in tons) of candy in a Yr 7 | 15 |
confectionery factory in Tanta city.
Discuss the possibility of forecasting profits using sales
Build a forecasting model
+Forecast the weekly profit resulting from sales of 10.3
tons
7/9/2023 Dr. Abdulfatah Salem