Chapter 9:

FOREC
ASTING
 Chapter Outline
9.1 BASIC CONCEPTS
9.2 SIMPLE MOVING AVERAGES
9.3 WEIGHTED MOVING AVERAGES
9.4 SIMPLE EXPONENTIAL SMOOTHING
9.5 ADJUSTED EXPONENTIAL
SMOOTHING
9.6 FORECAST RELIABILITY
9.7 SIMPLE LINEAR REGRESSION

2
 9.1
BASIC
CONCEPTS
 A FORECAST is a prediction, estimate, or
determination of what will occur in the future based on a
certain set of factors.

The value being

forecast may be sales,
interest rates, funds.
Gross national product
(GNP). technological
status, and others.

4
The factors on which a forecast is based
may be any of the following:

Past data Company data

Opinion or Perceived pattern

Judgment related to time

5
 “Major
categories of
forecasting
time
horizons”
6
 Time horizon
Intermediate- Long-term
Short-term Forecast
term Forecast Forecast
- It covers one day to - A period ranging - When a forecast
one year and are used from one season to covers from two to
mainly for short-run one or two years and five years or more
control such as is used for like market trends,
employment, production technology,
purchasing, schedules, revenues, facilities expansion,
scheduling, sales, and cash flow, and and general policy.
production rates. budget planning.

7
Forecastin
g
techniques
8
Qualitative Techniques Quantitative Techniques
- It is a statistical technique - It is a statistical technique
to make predictions about to make predictions about
the future which uses the future which uses
expert judgment instead numerical measures and
of numerical analysis. This prior effects to predict
method of forecasting future events. These
depends on the opinions techniques are based on
and knowledge of highly models of mathematics
qualified and experienced and in nature are mostly
employees to predict the objective. They are highly
future outcomes. dependent on
mathematical calculations.
9
 Qualitative Techniques
Executive Delphi Consumer Salesforce
Opinions technique surveys polling
In this method, panels In this method, the
The opinions of In this method, the
of experts are forecast is done based
experts from different survey is conducted
selected and are on the opinions of
departments are directly on the
individually salespeople who have
considered and customers on their
questioned about the steady interactions
averaged to forecast purchases. The
upcoming events. with the clients. As
the future sales. This surveys can be done
They do not form a they are closest to the
method of forecasting through telephone
group. For long-range customers, they can
can be done easily contacts, personal
forecasting, this better predict the
and quickly without interviews or
method is beneficial requirements of the
the necessity of questionnaires to
and very effective customers for the
elaborate statistics. obtain data from the
customers. future market.

It’s main disadvantage is

The main disadvantage of This method requires The main disadvantage of
that it depends on
this method is that from the extensive statistical this method is that the
individual opinions that may
returns there is lack of and analysis to test regarding salespeople can be either
not be unanimous and can
low reliability. the consumer behavior. optimistic or pessimistic
vary from individual to
about their predictions and
individual which could lead
this could lead to inaccurate
to wrong forecasting.
forecasting.
 These models examine the
past data patterns and forecast the
Time- future on the basis of underlying
series patterns that are obtained from
Quantitativ Analysis
those data. There are many types
of time series models like Simple

e and weighted moving average,

seasonal indexes, trend

Techniques projections, simple mean and

exponential smoothing.

Casual The model assumes that the

Methods variable that is being forecasted is
associated with other variables The
predictions are made based on
these associations. It is a method
that define as relationship among
independent and dependent
11 variables in a system of related
equations.
 Factors in
forecasting

12
 Trend - it is the general
movement or direction in the
Trend data.

Seasonal Factors -
Forecasting factors in which
variations in a time series
associated with a particular
Rando period of time.
m Factors Seasona
l Factors
Cyclical Factors - Forecasting
Factors factors applicable in longer-term
regular fluctuations which make
take several years to complete

Random Factors - These are

Cyclica event or effects that cannot be
l predicted with certainty but can
impact on the data.
Factors
 methods
(Time Series
Methods)
✘ Simple Moving Average
✘ Weighted Moving Average
✘ Simple Exponential
Smoothing
✘ Adjusted Exponential
Smoothing
✘ Forecast Reliability
14
Forecasting
important?

15
 9.2
Simple
Moving
Averages
 Simple ✘ The moving average

Moving model uses the last

periods in order to predict
Average demand in period.
✘ The moving average
model assumption is that
the most accurate
prediction of future
demand is a simple
(linear) combination of
17
past demand.
 Simple Moving Average
Formula:
Simple Moving
Average=

Or:

18
mple
:
Kroger sells bottled spring water for:

19
20
 9.3
Weighted
Moving
averages
Weighted Moving Average
● Is a time series forecasting method in
which the most recent data are weighted
heavier compared to the later data.
● This is desirable to vary the weights given
to historical data forecast future demand
or sales.

22
Weighted Moving
Average Formula:

23
 mple
:
The WSS motorcycle dealer in Quezon Avenue area wants
to accurately forecast the demand for the WSS hybrid
motorcycle during the next month. Because the distributor is
in Germany, it is difficult to send motorcycle back or
recorded if the proper number of motorcycles is not ordered
in a month ahead. From sales records, the dealer has
accumulated the following data for the past 11 months.
Determine the weighted averages forecast on demand with
the following weights:

a.) 20%, 30%, 50%

b.) 15%, 25%, 60%

24
25
26
27
 9.4
Simple
exponential
Smoothing
Simple Exponential
Smoothing
Refers to family of
forecasting models
that are very similar to
the weighted moving
average that weighs
the most recent past
data.

29
Determines the Typically not specific The value of α is
degree of smoothing and is usually done between 0.00 and
that takes place and by trial and error. 1.00
how responsive the
model is to
fluctuations of the
variable being
forecast.

30
Formula:
Ft+1 = αyt + (1- α) Ft
where:
Ft+1 = Forecast for the next period
yt = Actual data in the present
Ft = Previously determined forecast for the present
period
α = Smoothing constant
Or

Forecast = α(last value) + (1 – α)(last

forecast)

31
Example:
The WWS motorcycle dealer in Quezon Avenue area wants to accurately
forecast the demand for the WSS hybrid motorcycle during the next
month. Because the distributor is in Germany, it is difficult to send
motorcycle back or recorded if the proper number of motorcycles is not
ordered a month ahead. From sales records, the dealer has accumulated
the following data for the past 11 months. Establish the forecasts using
simple exponential smoothing if
α = 0.10 and α = 0.30.

Month Jan Feb Mar Apr May Jun July Aug Sept Oct Nov
Motorcycles
60 70 50 90 10 80 150 70 110 150 130

32
Month Jan Feb Mar Apr May Jun July Aug Sept Oct Nov

Motorcycles 60 70 50 90 10 80 150 70 110 150 130

Solution:

33
34
 Summary
Month Actual Forecast, Ft+1 Table for the
α = 0.10 α = 0.30
Simple
January 60 -- --
Exponential
Smoothing
February 70 60.00 60.00

March 50 61.00 63.00

April 90 59.90 59.10 Forecasts

May 10 62.91 68.37 with a = 0.10
June 80 57.62 50.86 and a = 0.30
July 150 59.86 59.60

August 70 68.87 86.72

September 110 68.98 81.70

October 150 73.08 90.19

November 130 80.77 108.13

December -- 85.69 35 114.69

36
 9.5
Adjusted
Exponential
Smoothing
 Exponential
Smoothing
It consists of simple The value of ß is Both α and ß is often
exponential between 0.00 to 1.00 determine
smoothing forecast similar to α. subjectively based
with trend on the judgment of
adjustment factor the forecaster.
added to it.

38
where:
AFt+1 = Adjusted exponential smoothing in the
present period
T = An exponential smoothed trendy factor.
Tt = The last period trend factor.
ß = Exponential smoothed trend for time period t.

39
Example:
The WWS motorcycle dealer in Quezon Avenue area wants to accurately
forecast the demand for the WSS hybrid motorcycle during the next
month. Because the distributor is in Germany, it is difficult to send
motorcycle back or recorded if the proper number of motorcycles is not
ordered a month ahead. From sales records, the dealer has accumulated
the following data for the past 11 months. Establish the forecasts using
simple exponential smoothing if α = 0.10 and α = 0.30.

Month Jan Feb Mar Apr May Jun July Aug Sept Oct Nov
Motorcycles
60 70 50 90 10 80 150 70 110 150 130

40
Month Jan Feb Mar Apr May Jun July Aug Sept Oct Nov

Motorcycles 60 70 50 90 10 80 150 70 110 150 130

Solution:

41
42
 Month Actual Adjusted Forecast
(ß = 0.20)

January 60 --

February 70 60.00
Summary
Table for the
March 50 61.80

April 90 59.66

May 10 65.11 Simple

June 80 55.14 Exponential
July 150 59.66 Smoothing
August 70 75.91
Forecasts
September 110 74.70
with a =
0.10 and a =
October 150 80.42

November 130 93.21

December -- 99.57 0.30

43
44
 9.6
Forecast
Reliability
 Forecast
Reliability
✘ It measures the closeness of
forecast in reflecting reality. And,
it is known that no forecast
measure will result in a
consistently perfect forecast.
✘ The Primary objective of forecasting is
to make it generally reliable or to see if
it is accurately portraying what
actually occurs.
46
 MEAN ABSOlutE
Deviation (MAD)
It is the average of the absolute value, or the
difference between actual values and their
average value, and is used for the calculation of
demand variability. It is expressed by the
following formula:

47
Example:
Example:
The WWS motorcycle dealer in Quezon Avenue area wants to accurately
forecast the demand for the WSS hybrid motorcycle during the next
month. Because the distributor is in Germany, it is difficult to send
motorcycle back or recorded if the proper number of motorcycles is not
ordered a month ahead. From sales records, the dealer has accumulated
the following data for the past 11 months. Establish the forecasts using
simple exponential smoothing if α = 0.10 and α = 0.30.

Month Jan Feb Mar Apr May Jun July Aug Sept Oct Nov
Motorcycles
60 70 50 90 10 80 150 70 110 150 130

49
49
Mean Absolute Deviation
Values of all the
forecasts:

3-month Moving Average=

36.38
5-month Moving Average=
39.17
Weighted Moving
Average(20%,30%,50%)=
40.13
Weighted Moving
Average(15%,25%,60%)=
42.19
Simple Exponential
Smoothing(α=0.10)=38.5
7
Simple Exponential
Smoothing(α=0.30)=35.9
5
Adjusted Exponential 50
Smoothing(α=0.20)=37.1
 9.7
Simple Linear
Regression
 Regression
Analysis
Is a simple statistical tool use to
model the dependence of a
variable one (or more)
explanatory variables.
52
 Simple
Linear Is the least estimator of a linear regression model with a

Regress
Least
single predictor (or one independent variable).

ion
square
Determines a regression equation by minimizing the
sum squares of the vertical distances between the
actual y values and the predicting values of y.
model
Resid The difference between an observed and
predicted value. The mean of a residual is

ual always zero.

Outl The difference between an observed and

predicted value. The mean of a residual is

iers always zero.

Influentia Scores whose removal greatly changes the

l scores regression

53
 ASSUMPTIONS OF LINEAR
REGRESSION EQUATION
LINEARITY NORMALLY
The mean of each error DISTRIBUTED
component is 0
ERROR TERMS
Each error component
(random variable) follows an
approximate normal
distribution.
INDEPENDENCE OF HOMOSCEDASTICIT
ERROR TERMS Y
The errors are independent of The variance of the error
each other components is the same for
each value of the independent
variable

54
Estimat
ing the Ŷ = predicted or fitted value of y

Coeffici X = the value of any particular

observations of the independent variable

ent Y = the value of any particular

observations of the dependent variable
𝑏_1 = slope of the regression line
𝑏_0 = intercept of the regression line
x̄ = mean of independent variable
ȳ = mean of the dependent variable

55
 Sum of
✘ SSE is the sum of Squares
the
squared differences error
between each observation
and its group's mean.
✘ It can be used as a
measure of variation
within a cluster. If all
cases within a cluster are
identical the SSE would
then be equal to 0. 56
Sum of squares for
error

57
 Standard
error of
Estimate
Is the standard deviation
of the observed y values
about the predicted y
58
 Error
of
𝑠_𝐸= standard error of estimate
Estimat
𝑏_0 = intercept of regression line
𝑏_1 = slope of regression line
X = the value of any particular
e
observations of the independent
variable
Y= the value of any particular
observations of the dependent variable
Ŷ = predicted or fitted value of y
N = the sample size
59
 A Coefficient of non-
determination is the
Is the measure of proportion in the independent
variation of the variable that is left unexplained
by the independent variable,
dependent variable determined by 1-r^2
that is explained by the
regression line and the
independent variable

Coefficient 60
Coeffici
ent of
determi
nation

61
 Confidence
A confidence
interval is a range Interval
of values Represents a closed
associated with a
interval where a certain
population
parameter. For percentage of the
example, the population is likely to lie.
mean of a
population.
Prediction
A prediction
interval is where Interval
you expect a
it is a range of values
future value to
fall.
that predicts the value of
62
Confidence
intervals are always
associated with a
confidence level,
representing a degree of
uncertainty (data is random,
and so results from
statistical analysis are never
100% certain).
For example, you might say that the
mean life of a battery (at a 95%
confidence level) is 100 to 110
hours. This tells you that a battery
will fall into the range of 100 to 110
hours 95% of the time. 63
Predictive
intervals
tells you where a value
will fall in the future, given
enough samples, a certain
percentage of the time.

For example, A 95% prediction

interval of 100 to 110 hours for
the mean life of a battery tells
you that future batteries
produced will fall into that range
95% of the time. There is a 5%
chance that a battery will not fall
into this interval.
64
Confidence Interval
and prediction
Interval

65
 What
remains to
be
resolved
is the
question
of
knowing
to what
extent
and up to
66
what point
Presented by:
Group 9
Cruz, Gian Carlo
Mampusti, Charmaine
Miguel, Demie
Reynoso, Jolina
Salayo, Dharlene Jeanne
Villanueva, Tommy
67
THANKS!