S6 - Time - Series Analysis - 1

Time – Series
Analysis
[1] CHAPTER 14: TIME-SERIES ANALYSIS
• 1. Time-Series Components
• 2. Trend Forecasting
Contents • 3. Assessing Fit
• 4. Moving Averages
• 5. Exponential Smoothing
• 6. Seasonality
LEARNING OBJECTIVES
• Define time-series data and its components.
• Interpret a linear, exponential, or quadratic trend model.
• Fit any common trend model and use it to make forecasts.
• Know the definitions of common fit measures.
• Interpret a moving average and use Excel to create it.
• Use exponential smoothing to forecast trendless data.
• Interpret seasonal factors and use them to make forecasts.
• Use regression with seasonal binaries to make forecasts.
• Interpret index numbers.
1. Time-Series Components
• A time-series variable (denoted Y) consists of data observed over n periods of time.

• Businesses use time-series data
o to monitor a process to determine if it is stable.
o to predict the future (forecasting).
• Time-series data can also be used to understand economic, population, health, crime,
sports, and social problems.
• Time-series graph:
• Time-series data are usually plotted as a line graph,
• time is on the horizontal axis and the variable of interest (𝑌) on the
vertical axis → to reveals how a variable changes over time.
• Fluctuations are easier to see on a line graph.
• Notation for time series data:
• 𝑌𝑡 is the value of the time series in period 𝑡.
• 𝑡 is an index denoting the time period (𝑡 = 1, 2, 3, … , 𝑛) .
• 𝑛 is the number of time periods.
• 𝑌1 , 𝑌2 , 𝑌3 , … , 𝑌𝑛 is the data set for analysis
• The periodicity is the time interval over which data are collected
(decade, year, quarter, month, week, day, hour).
1. Time-Series
Components
• Time-series decomposition
seeks to separate a time-series
Y into four components:
• Trend (T )
• Cycle (C )
• Seasonal (S )
• Irregular (I ).
• These components are
assumed to follow either an
additive or a multiplicative
model.
Additive vs Multiplicative models
• The additive form is attractive for its simplicity.

• The multiplicative model is often more useful for forecasting financial data, particularly
when the data vary over a range of magnitudes.
• The multiplicative model becomes additive if logarithms are taken (data are nonnegative):
log 𝑌 = log 𝑇 × 𝐶 × 𝑆 × 𝐼 = log 𝑇 + log 𝐶 + log 𝑆 + log(𝐼)
• Trend
• Trend (𝑇) is the general movement over all years (𝑡 = 1, 2, … , 𝑛).

• Trends may be steady and predictable, increasing, decreasing, or staying the same.
• A mathematical trend can be fitted to any data but may or may not be useful for predictions.
Cycle
• Cycle (𝐶) is a repetitive up-and-down
movement around the trend that covers
several years.
• Over a small number of time periods (a
typical forecasting situation), cycles are
undetectable or may resemble a trend. For
this reason cycles are not discussed further
in this chapter.
Seasonal
• Seasonal (𝑆) is a repetitive cyclical pattern
within a year (or within a week, day, or
other time period).
• By definition, annual data have no
seasonality.
Irregular
• Irregular (I) is a random disturbance
that follows no pattern.
• It is also called the error component
or random noise reflecting all
factors other than trend, cycle, and
seasonality.
2. Trend Forecasting
main categories of forecasting models
Three trend models
• There are many possible trend models, but three of them are especially
useful in business:
• 𝑌𝑡 = 𝑎 + 𝑏𝑡 𝑓𝑜𝑟 𝑡 = 1,2, … . , 𝑛 (Linear trend)
• 𝑌𝑡 = 𝑎. 𝑒 𝑏𝑡 𝑓𝑜𝑟 𝑡 = 1,2, … . , 𝑛 (Eponential trend)
• 𝑌𝑡 = 𝑎 + 𝑏𝑡 + ct 2 𝑓𝑜𝑟 𝑡 = 1,2, … . , 𝑛 (Quadratic trend)
• The linear and exponential models are widely used because they have
only two parameters and are familiar to most business audiences.
• The quadratic model may be useful when the data have a turning point.
• All three can be fitted by Excel, R.
2. Trend Linear trend model
• The linear trend model (𝑌𝑡 = 𝑎 + 𝑏𝑡) is useful for a
Forecasting time-series that grows or declines by the same
amount (𝑏) in each period.
2. Trend Linear trend model: Example
Forecasting • Use Price data to forecast for the next 3 years.
Year Price Year Price Year Price Year Price

1995 105 2002 156 2009 172 2016 182
1996 117 2003 150 2010 167 2017 177
1997 131 2004 163 2011 170 2018 179
1998 131 2005 157 2012 169 2019 184
1999 143 2006 160 2013 173 2020 ?
2000 145 2007 166 2014 171 2021 ?
2001 153 2008 165 2015 181 2022 ?
2. Trend Exponential trend model
• The exponential trend model (𝑌𝑡 = 𝑎. 𝑒 𝑏𝑡 ) is useful for
Forecasting a time series that grows or declines at the same rate
(𝑏) in each period.
When to Use the Exponential Model
• Financial data or data that cover a longer
period of time. (Financial analysts: costs,
revenue, and salaries are best projected
under assumed percent growth rates.)
• compare two growth rates in two time-
series variables with dissimilar data
units.
• There may not be much difference
between a linear and exponential model
when the growth rate is small and the
data set covers only a few time periods.
Estimate Exponential trend model
• Transform the exponential trend model 𝑌𝑡 = 𝑎. 𝑒 𝑏𝑡 by taking logarithm two sides of
the equation:
log 𝑌𝑡 = log 𝑎. 𝑒 𝑏𝑡 = log 𝑎 + 𝑏𝑡
Example
• Apply Exponential trend model to estimate and forecast for Price data.
2. Trend
Forecasting
Quadratic Trend Model
• The quadratic trend model
(𝑌𝑡 = 𝑎 + 𝑏𝑡 + c. 𝑡 2 , the 𝑡 2
term allows a nonlinear
shape) is useful for a time
series that has a turning
point or that is not
captured by the
exponential trend model.
• Example: Price data
2. Trend
Forecasting
• You can increase the 𝑅2
by choosing a more
complex model. But if you
are making a forecast, this
is not the only relevant
issue because 𝑅2
measures the fit to the
past data.
3. Assessing Fit

S6 - Time - Series Analysis - 1

Uploaded by

Copyright:

Available Formats

S6 - Time - Series Analysis - 1

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

S6 - Time - Series Analysis - 1

Uploaded by

Copyright:

Available Formats

Time – Series

• A time-series variable (denoted Y) consists of data observed over n periods of time.

• The additive form is attractive for its simplicity.

• Trend (𝑇) is the general movement over all years (𝑡 = 1, 2, … , 𝑛).

Forecasting • Use Price data to forecast for the next 3 years.

Year Price Year Price Year Price Year Price

You might also like