Introduction

Time-Series Data: A time series is a set of measurements,

ordered over time, on a particular quantity of interest.
• Numerical data ordered over time
• The time intervals can be annually, quarterly, daily, hourly,
etc.
• The sequence of the observations is important
• Example:

Year: 2008 2009 2010 2011 2012

Sales: 75.3 74.2 78.5 79.7 80.2
Copyright © 2013, 2010, 2007 Pearson Education, Inc. All Rights Reserved Slide - 1
 The Importance of Forecasting

The Importance of Forecasting

 Government needs to forecast unemployment,
interest rates, expected revenues from income taxes
to formulate policies
 Marketing executives need to forecast demand, sales,
consumer preferences in strategic planning
 College administrators need to forecast enrollments
to plan for facilities and for faculty recruitment
 Retail stores need to forecast demand to control
inventory levels, hire employees and provide training

Copyright © 2013, 2010, 2007 Pearson Education, Inc. All Rights Reserved Slide - 2
Time-Series Plot
A time-series plot is a two-dimensional plot of
time-series data

• the vertical axis

measures the variable
of interest
• the horizontal axis
corresponds to the
time periods

Copyright © 2013, 2010, 2007 Pearson Education, Inc. All Rights Reserved Slide - 3
Trend Component (1 of 2)
• Long-run increase or decrease over time
(overall upward or downward movement)
• Data taken over a long period of time

Copyright © 2013, 2010, 2007 Pearson Education, Inc. All Rights Reserved Slide - 4
Trend Component (2 of 2)
• Trend can be upward or downward
• Trend can be linear or non-linear

Copyright © 2013, 2010, 2007 Pearson Education, Inc. All Rights Reserved Slide - 5
Seasonal Component
• Short-term regular wave-like patterns
• Observed within 1 year
• Often monthly or quarterly

Copyright © 2013, 2010, 2007 Pearson Education, Inc. All Rights Reserved Slide - 6
Cyclical Component
• Long-term wave-like patterns
• Regularly occur but may vary in length
• Often measured peak to peak or trough to trough

Copyright © 2013, 2010, 2007 Pearson Education, Inc. All Rights Reserved Slide - 7
Irregular Component
• Unpredictable, random, “residual” fluctuations
• Due to random variations of
– Nature
– Accidents or unusual events
• “Noise” in the time series

Copyright © 2013, 2010, 2007 Pearson Education, Inc. All Rights Reserved Slide - 8
 Time Series Forecasts
Uses time-ordered sequence of observations at regular intervals

 Trend - long-term movement in data

 Seasonality - short-term regular
variations in data
 Cycle – wavelike variations of more than
one year’s duration
 Irregular variations - caused by unusual
circumstances
 Random variations - caused by chance

3-9
 Techniques for Averaging

 Moving average
 Weighted moving average
 Exponential smoothing

3-10
 Moving Averages
 Moving average – A technique that averages a
number of recent actual values, updated as
new values become available.
At-n + … At-2 + At-1
Ft = MAn=
n
 Weighted moving average – More recent
values in a series are given more weight in
computing the forecast.
wnAt-n + … wn-1At-2 + w1At-1
Ft = WMAn=
n
3-11
 Moving Average Example
Calculate a three-period moving average forecast for demand in period 6

If the actual demand in period 6 is 38, then the moving average forecast for
period 7 is:

3-12
 Simple Moving Average
Actual
MA5
47
45
43
41
39
37 MA3
35
1 2 3 4 5 6 7 8 9 10 11 12

Takeaways:
• Fewer data points (e.g., the 3 month moving average) is more sensitive
to real life and is a more dynamic forecast.
• Larger data points (e.g., the 5 month moving average) is smoother (less
reactionary)
3-13
Weighted Moving Average Example
Period Demand Weight

1 42

2 40 10%

3 43 20%

4 40 30%

5 41 40%

Takeaways:
• Choice of weights must add up to 100%
• Choice of weights are discretionary
• Choice of weights are often based on trial and error, and forecaster’s
experience.
3-14
 Exponential Smoothing

Ft = Ft-1 + (At-1 - Ft-1)

• 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.
 Uses most recent period’s actual and forecast data
 Weighted averaging method based on previous
forecast plus a percentage of the forecast error
 A-F is the error term,  is the % feedback

3-15
Exponential Smoothing Example

3-16
Exponential Smoothing Example
Period 2
forecast
is a naïve
forecast

F3  F2   ( A2  F2 )
  0.10, F3  42  0.10(40  42)  42  0.10  ( 2)  42  0.2  41.8

F4  F3   ( A3  F3 )
  0.10, F4  41.8  0.10(43  41.8)  41.8  0.10 1.2  41.8  0.12  41.92

3-17
 Picking a Smoothing Constant

Takeaway:
• Low α, less sensitive to forecast error, relatively smoother (stable) forecast
• Large α, more sensitive to forecast error, relatively more dynamic forecast

3-18
 Linear Trend Equation
Ft

Ft = a + bt

0 1 2 3 4 5 t
 Ft = Forecast for period t
 t = Specified number of time periods
 a = Value of Ft at t = 0
 b = Slope of the line

3-19
Calculating a and b

n  (ty) -  t  y
b =
n t 2 - (  t) 2

 y - b t
a =
n

3-20
 Time-Series:
Linear Trend Example
Period Demand
(Week) (Sales/Week)
1 150
2 157
3 162
4 166
5 177

3-21
Linear Trend Equation Example
t y
2
Week t Sales ty
1 1 150 150
2 4 157 314
3 9 162 486
4 16 166 664
5 25 177 885

2
 t = 15 t = 55  y = 812  ty = 2499
2
(t) = 225

3-22
 Linear Trend Calculation
5 (2499) - 15(812) 12495-12180
b = = = 6.3
5(55) - 225 275 -225

812 - 6.3(15)
a = = 143.5
5

y = 143.5 + 6.3t

3-23