CHAPTER V.

FORECASTING
Forecasting Components
Forecasting is the process of making future assumptions by analyzing trends, including
historical and present data. It is the first step in the planning and scheduling process for
most goods and service organizations.
A variety of forecasting methods exist, and their applicability is dependent on the time
frame of the forecast (i.e., how far in the future we are forecasting), the existence of
patterns in the forecast (i.e., seasonal trends, peak periods), and the number of variables to
which the forecast is related.
In general, forecasts can be classified according to three time frames: short range,
medium range, and long range.
Short-range forecasts. Typically encompass the immediate future and are concerned with
the daily operations of a business firm, such as daily demand or resource requirements.
These also include decisions covering months to a year at most such as purchasing and
deployment options.
Medium-range forecasts. Typically encompasses anywhere from 1 or 2 months to 1 year. A
forecast of this length is generally more closely related to a yearly production planning and
budgeting and will reflect such items as peaks and valleys in demand and the necessity to
secure additional resources for the upcoming year.
Long-range forecasts. Typically encompasses a period longer than 1 or 2 years. Long-
range forecasts are related to management's attempt to develop new products for changing
markets, expansion of facilities, or secure long- term financing. In general, the further into the
future one seeks to predict, the more difficult forecasting becomes.
Forecasting Approaches
In general, there are two approaches to forecasting: qualitative and
quantitative (Shim et al, 2006)

1. Qualitative Forecasting:
It applies when historical data are unavailable or very little. Qualitative forecasting
methods are considered to be highly subjective and judgmental. Managers use
judgement, intuition, knowledge and skill to make effective forecasts.

The qualitative forecasting methods are as follows:

a. Jury of Executive Opinion

b. Delphi method

c. Sales force polling

d. Consumer market surveys

Jury of Executive Opinion
• Under this method, the view of the top managers or a group with
a high level of expertise, often in combination with statistical
models, are synthesized to generate a consensual forecast.

• The opinions of experts become the basis of forecasting. They

are generally most familiar with their firms' own capabilities and
resources and the markets for their products.

• The disadvantage of this method is the presence of bias. Since

the experts are in the same field or group, their mentality or way
of thinking is homogenous.
Delphi Method
• The Delphi method is a procedure for acquiring informed judgments and
opinions from knowledgeable individuals, using a series of questionnaires to
develop a consensus forecast about what will occur in the future.

• Responses acquired from the experts are analyzed by an independent party

that will provide the feedback to the panel members. Based on the responses
of other individuals, each expert is then asked to make a revised forecast.

• This process continues till a consensus is reached or until further iterations

generate no change in estimates.

• This method has been especially useful for forecasting technological change
and advances.
Sales Force Polling
• Under “Sales-force polling” or “Opinion poll method”, sales representatives,
professional experts, the market consultants and others are asked to express their
considered opinions about the volume of sales expected in the future. The logic and
reasoning behind the method is that the sales force is a direct point of contact with
the consumer.
• This contact provides an awareness of consumer expectations in the future that
others may not possess. Salesmen, being very close to the customers, will be in a
position to know and feel the customers’ reactions towards the product. They can
study the pulse of the people and identify the specific views of the customers.
• These people are quite capable of estimating the likely demand for the
products with the help of their intimate and friendly contact with the
customers and their personal judgments based on the past experience.
• Thus, they provide approximate, if not accurate estimates. Then, the
views of all salesmen are aggregated to get the overall probable demand
for a product.
Consumer Market Survey
• Consumer market survey is an organized approach that uses
surveys and other research techniques to determine what products
and services customers want and will purchase, and to identify
new markets and sources of customers.

• Surveys are conducted through telephone inquiries,

questionnaires, and interviews. Surveys can help not only in
preparing a forecast but also in improving product design, planning
for new products, and determining consumer behavior.
2. Quantitative Forecasting:

Quantitative forecasting methods make use of historical data. The goal of these
methods is to use past data to predict future values.

Quantitative forecasting methods are subdivided into two types: time series and
causal.

• Time-series forecasting methods involve forecasting future values based

entirely on the past and present values of a variable.
• Causal forecasting methods involve the determination of factors that relate to
the variable you are trying to forecast.
Time Series Forecasting
• Time series forecasting assumes that the future is a function of the past.
Thus, historical data are used to predict the future using sequences with
equal periods. The sequence may be daily, weekly, monthly, quarterly,
annually or any equal periods the firm may think of.

• If the yearly sales of a certain product are to be predicted, the past

year’s sales will be used in making the forecasts. If the firm is
forecasting monthly sales, then past monthly data are used.

• Forecasting time-series data that future values are predicted only by

using past values. Other variables, no matter how significant, are
ignored.
Decomposition of a Time Series Forecast

Analyzing a time series means breaking down past data into

components and then projecting them forward. A time series typically
has four components, trend, seasonality, cycle, and random variations.

1. Trend is the gradual upward or downward movement of the data

over time. Changes in income, population, age distribution, or cultural
views may account for movement in trends.
2. Seasonality is a data pattern that repeats itself after a period of
days, weeks, months, or quarters.
3. Cycle is a pattern of the data that varies in length and occurs every
several years usually 2 to 10 years. It is usually associated with the
business cycle and is very important in short-term business analysis and
planning. A business cycle is difficult to predict because of armed
conflicts, political upheavals, economic development, and social changes
in global scene.
4. Random variations are “blips” in the data caused by chance and
unusual situations. They follow no discernable pattern so they cannot be
predicted.
Quantitative Forecasting Models

The quantitative forecasting models under time series of

forecasting are:

a. Moving average
b. Weighted moving average

c. Naïve model
d. Exponential smoothing
A moving average is a
technique that calculates the
overall trend in a data set. In
operations management, the
data set is sales volume from
historical data of the
company. This technique is
very useful for forecasting
short-term trends. It is
OBJECTIVES

• CALCULATE MOVING AVERAGES

• COMPUTE ERROR MEASURES

 MAD-MEAN ABSOLUTE DEVIATION 0R (MAE)

 MSE-MEAN SQUARED ERROR
 MAPE-MEAN ABSOLUTE PERCENT ERROR
 We will be calculating 3 week moving averages and also
compute a moving average forecast, and also compute a
moving average forecast for week 8.
 We start by calculating the moving average for the 4th week. That is
we average the sales values for the first 3 weeks to produce a
moving average forecast for week 4.
 Next we calculate the forecast errors. Forecast errors are
calculated as Actual Values- corresponding forecast values.
 Since week 1 to 3 have no forecast values, we begin calculating
errors in week 4.
 Next we calculate the Mean Absolute Deviation or Error
 Note that Error and Deviation are used interchangeably. The two bars ( I I) around Error denotes
absolute values. That is, return a positive value whether the original value is positive or negative.

 To find the Mean Absolute Deviation (MAD or MAE), we simply average these absolute errors.
 First, we sum up the absolute errors, and then divide the total by 4 since there are 4 errors.
 Next we calculate the Mean Squared Error. In this case, we first square the
errors and then average them .
 Since the square of negative values is always positive, we can simply square the
absolute errors
 Finally, let us compute the Mean Absolute Percent Error
 The absolute percent error is a measure of the error as a percentage of actual values
 To obtain the absolute percent error, we simply divide the absolute errors by the actual sales,
and multiply by 100
Weighted Moving Average
• A weighted moving average (WMA) is more powerful and
economical compared with moving average. When a detectable
trend or pattern is present, weights can be used to place more
emphasis on recent values.
• This practice makes the forecasting technique more responsive to
change because the more recent periods are heavily weighted.
• Choice of weight is somewhat arbitrary because there is no set
formula to determine them.
• Therefore, deciding which weight to use requires some experience.
It should be noted that the sum of weights should always equal to
one.
The weighted moving average (WMA) is a technical
indicator that assigns a greater weighting to the most
recent data points, and less weighting to data points in
the distant past. The WMA is obtained by multiplying
each number in the data set by a predetermined
weight and summing up the resulting values.
OBJECTIVES

• CALCULATE WEIGHTED MOVING AVERAGES

• COMPARE ERROR MEASURES USING THE MEAN
ABSOLUTE DEVIATION (MAD)
EXAMPLE

Forecast sales using 4-

week weighted moving
averages with weights
0.4, 0.3, 0.2, and 0.1
• In practice, the weighted moving average is usually
employed when there is a need to place more importance
on some periods over the others.
• In most cases, we place more importance on more recent
data

• Therefore, in this example, the 0.4 weight will be placed on

the most recent value, the 0.3 on the next most recent, and
so on.
• Let’s now calculate 4-week weighted moving averages
using the given weights; .4, .3. .2, and .1

• Since we’re computing a 4-week moving averages, we start

by using data from the first 4 week to compute the moving
average forecast for week 5
• Next, we calculate the Mean Absolute Deviation
• First, we calculate the Absolute errors. That is, the positive
difference bet ween the actual and forecast values and
then, average them.
• Now note, the weights on this
first example 0.4, 0.3, 0.2, and
0.1 added up to 1

• Let’s look at the next example

where the weights do not add up
to 1
EXAMPLE

Forecast sales using 2-

week weighted moving
averages with weights
3 and 2
weights 3 and 2 = 5
• In calculating the weighted moving averages,
we multiply the sales values by the weights as
we did before, but in this case, we also divide
by the total weights which is 5
 • Next, we calculate the mean absolute deviation

 Note that Error and Deviation are used interchangeably. The two bars ( I I) around Error denotes
absolute values. That is, return a positive value whether the original value is positive or negative.
 Comparing Forecast Errors

• Since the MAD is an error measure, smaller

MADs produce better smoothing of the data.
• Therefore, using MAD, the 2-week weighted
average method produced a better forecast
Naïve Model

• The naïve method of forecasting dictates that we use the previous

period to forecast for the next period.
• A model in which minimum amounts of effort and manipulation of
data are used to prepare a forecast.
• Most often naïve models used are random walk (current value as
a forecast of the next period) and seasonal random walk (value
from the same period of prior year as a forecast for the same
period of forecasted year.)
Exponent Smoothing
• This method uses a series of exponentially weighted moving
averages techniques where more weight is given to the recent
data.
• The weights assigned to the values change so that the most
recent value receives the highest weight, the previous value
receives the second-highest weight, and so on, with the first
value receiving the lowest weight.
• Throughout the series, each exponentially smoothed value is
supported by the belief that the future is more dependent on
the recent past than on the distant past.
• Exponential smoothing is a popular technique that does not
involve voluminous records to forecast, and is easy to use and
effective for short-run forecasting. This method is known to be
useful on random historical data with no seasonal fluctuations.
OBJECTTIVES
• Calculate Forecast using Exponential Smoothing
• Calculate Mean Squared Error MSE
• We will be calculating
exponential smoothed
forecast for these sales
values collected over a 7-
week period.
• The actual sales will be
represented by At. Some
authors used Yt to
represent actual values.
• Forecasts values on the
other hand are generally
represented by Ft
• The Exponential Smoothing Method uses the
formula
• Ft +1 = Ft + a (At-Ft) where alpha is a value
between 0 and 1, referred to as the smoothing
constant
• These means that the forecast for the current
period is obtained by adding the forecast in
the last period to a fraction of the error from
the last period
• To make the calculation easier, this formula
can be rewritten as

• We will be using this second formula for our

calculations.
• Note that both formula will give the same
result. It is just a little easier to use the
second one
• Our first objective is to
calculate exponential using
α=0.2
• Since α=0.2, the 1 minus alpha will be1-0.2, which is equals
to 0.8
• Then the formula becomes:
• Since the forecast requires both actual and forecast values
from the last period, we are sometimes given a forecast value
for the first period.
• If no forecast value for the first value is given , then we assume
F1=A1.
• That is the first forecast value is assumed to be the first actual
value.
• So F1= 39.
• We then calculate F2 as 0.2 (39) + 0.8 (39), which will again
give the same value.
• As a result, we just usually assume that F2=A1 and not bother
calculating it at all
• Next we compute the mean squared error
• To obtain the mean squared error MSE, we first obtain the
forecast errors, square them, and then find the mean or
average of the squared errors
• When computing errors for exponential smoothing
forecasts, we do not calculate and error for period 1 unless
otherwise stated.
• So the error for week 2 is 44-39=5
• Next we squared the errors, starting with week 2
• Since we only have 6 periods with errors, we add up these
squared errors and divide by 6 to obtain a Mean Squared
Error of 11.58