OTHM Level 7

Supply Chain Planning, Modelling

and Analytics

Forecasting Techniques
Time Series and Regression Analysis

Sidath Waidyasekera
MBA(PIM-USJ), PG Dip Mkt(UK), MCIM(UK), MILT(UK),
MIDPM(UK), MSLIM(SL), MIM(SL)
Chartered Marketer

Winfield Academy of Business

& Finance
1
Forecasting Techniques
• Qualitative approaches to forecasting,
• Quantitative approaches to forecasting,
• The components of a time series ,
• Using smoothing model of the forecasting,
• Measures of forecasting accuracy,
• Using trend projection in forecasting,
• Using regression analysis in forecasting.
 Forecasting Introduction
• An essential aspect of managing any organization is
planning for the future,
• Organizations employ forecasting techniques to
determine future inventory , cost, capacities and
interest rate changers,
• There are two basic approaches to forecasting ,
Qualitative
Quantitative
 Qualitative Approaches to
Forecasting
 Delphi approach
• A panel of experts, each of whom is physically separated
from the others and is anonymous, is asked to respond to
a sequential series of questionnaires.
• After each questionnaire, the responses are tabulated
and the information and the opinions of the entire group
are made known to each of the other panel members so
that they may revise their previous forecast response.
• The process continues until some degree of consenses
is achieved.
Qualitative Approaches (continued)
 Scenario writing
• Scenario writing consists of developing a
conceptual scenario of the future based on a well
defined set of assumptions.
• After several different scenarios have
been developed, the decision maker determines
which is most likely to occur in the future and
makes decisions accordingly.
Qualitative Approaches (continued)
 Subjective or interactive approaches
• These techniques are often used by committees of
panels seeking to develop new ideas or solve complex
problems.
• They often involve “brainstorming sessions” .
• It is important in such sessions that any ideas or options
be permitted to be presented without regard its relevancy
and without fear of criticism.
 Quantitative Approaches to
Forecasting
 Quantitative methods are based on an analysis of
historical data concerning one or more time series.
 A time series is a set of observations measured at
successive points in time or over successive periods
of time.
 If the historical data used are restricted to past values
of the series that we are trying to forecast, the
procedure is called a casual method.
 Quantitative methods are generally preferred. In this
chapter we will focus on quantitative approaches to
forecasting.
 Time Series Data
 Time series data is usually plotted on a graph to
determine the various characteristics or components
of the time series data.
 There are 4 major components; Trend, Cyclical,
Seasonal, and irregular Components.
 Components of Time Series
 The trend component for the gradual shifting of the
time series over a long period of time.
 Any regular pattern of sequences of values above and
below the trend line is attributable to the cyclical
components of the series.
 The seasonal component of the series account for
regular patterns of variability within certain time
periods, such as over a year.
 The irregular component of the series is caused by
short term unanticipated and non recurring factors that
affect the values of the time series. One cannot
attempt to predict its impact on the time series in
advance.
 Time Series Data
 We will learn the following forecasting approaches.
• Smoothing
• Trend projections
Excel Instructions for Drawing a Scatter Plot

1. Enter data in the Excel spreadsheet.

2. Click on Insert on the toolbar and then click on the
chart tab. The Chart Wizard will appear. In step 1 on
select the XY (scatter) chart type and then click next.
3. In step 2 specify the cells where your data is located
in the data range box.
4. In step 3 you can give your chart a title and label
your axes. In step 4 specify where you want the
chart to placed.
 Smoothing Methods

 In cases in which the time series is fairly stable and

has no significant trend, seasonal or cyclical effects,
one can use smoothing methods to average out the
irregular components in the time series.

 Three common smoothing methods are,

1. Moving average
2. Weighted moving average
3. Exponential smoothing
Smoothing Methods: Moving Average

Moving Average Method

The moving average method consists of

computing an average of the most recent
n data values for the series and using this
average for forecasting the value of the time
series for the next period.
Robert Drug’s Example: Moving Average

Our scatter plot for Robert’s drug sales has no

significant trend, seasonal, or cyclical effects. Thus
we should employ a smoothing technique for
forecasting sales.

Forecasting the sales for period 11 using a three

period moving average (MA3)
 Smoothing methods: Weighted Moving
Average
• Weighted moving average method
The weighted Moving Average Method consists of computing a
weighted average of the most recent n data values for the
series and using this weighted average for forecasting the
value of the time series of the next period. The more recent
observations are typically given more weight than older
observations. For convenience, the weights usually sum to 1.
• The regular moving average gives equal weight to past data
values when computing a forecast for the next period. The
weighted moving average allows different weights to be allocated
to past data values.
There is no Excel command for computing this so you must do this manually. You can
either manually enter the formulas into excel and apply to all periods or compute value
by hand.
 Robert’s Drugs:
Exponential smoothing

• Forecast the sales for period 11 using Exponential

Smoothing α= 0.1
 Questions that you should be asked
• For the moving average technique, how do I determine the
best value of n to use for forecasting?
• For Exponential Smoothing, how do I determine the best
value of α to use?
• If I realize that a smoothing technique should be
employed,
how do you know which smoothing technique is the
best?
In order to answer the above questions, we need criteria for
judging the accuracy of a forecasting technique. Once we
select a criterion, the method (or parameter) which provides
the best value for are criterion is the best method (or
parameter) to use for forecasting our scenario.
Measures of Forecast Accuracy
• Mean Squared Error (MSE)
The average of the squared forecast errors for the historical
data is calculated. The forecasting method or parameter(s)
which minimize this mean squared error is then selected.
• Mean Absolute Deviation (MAD)
The mean of the absolute values of all forecast errors is
calculated, and the forecasting method or parameter (s) which
minimize this measure is selected. The mean absolute
deviation measure is less sensitive to individual large forecast
errors than the mean squared error measures.

You may choose either of the above criteria for evaluating the accuracy of a
method (Or parameter)
Selecting the best Smoothing Technique for
Robert’s Drugs
• Determine the smoothing technique that is best for
forecasting Robert’s drug sales. A two period moving
average, a three period moving average,
exponential smoothing (α=0.1) ,or exponential
smoothing (α=0.2)

Realistically we should have experimented with more values of

n for the moving average, and α for exponential smoothing to
determine the absolute best parameters to use for our
technique.

On the next side we randomly chose to use the MSE criterion to

judge the best technique.
Robert’s drugs; Comparing Smoothing
Techniques
• Since the three period moving average technique
(MA3) provides to lowest MSE value, this is the best
smoothing technique to use for forecasting
Robert’s Drug sales.
 Trend Projection
• If a time series exhibits a linear trend, the method of
least squares may be used to determine a trend
line (projection) for future forecasts.
• Least squares, also used in regression analysis,
determines the unique trend line forecast which
minimizes the mean square error between the trend
line forecasts and the actual observed values for the
time series.
• The independent variable is the time period and the
dependent variable is the actual observed value in
the time series.
 Casual method: Regression
Analysis
Regression analysis is similar to trend analysis, except
the independent variable is not restricted to time. Refer
to Robert’s Drug example. Instead of letting time
represent our independent variable, we could forecast
sales based upon the price of the product. Since
products often go on sale, we could collect data over
several months collecting the weekly price and number
of items sold for the week. For this model, we would find
the regression equation in the same manner in which
we found the trend line except we would call the
independent variable x, instead of t.
 Regression analysis in excel
• The dependent variable Y can predicted using the
same forecast function in excel as used to forecast a
trend line.
Forecast Accuracy - Product Code Level Last 12 Months
Last Month Orders Forecast
M+2 forecasted excess
- Demand vs Forecast = 94.0% (48.5k - excess) 89.06% 12.55%
- 28k & 13k forecast excess from Products…
- 10 each forecast excess from Products …
Orders Forecast
- 32.6k more demand for Products
M+1 forecasted excess
89.69% 12.17%

115% demand, 25k

& 24k more
Products ..

83.9% demand,
34k & 30k excess
Products …
 Forecast Accuracy
• Make sure to deeply analyze the product code
level – line graphs with last 18 months behavior ,
• Demand vs Forecast –Worst product code or
material ,
• Forecast Excess ,
• Demand vs forecast top level products ,
• Actual sales Vs Forecast.