Scribd
Upload
23 views29 pages

Multiplicative Seasonal ARIMA Models

Uploaded by

Reetika Choudhury
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
23 views29 pages

Multiplicative Seasonal ARIMA Models

Uploaded by

Reetika Choudhury
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 29

Multiplicative Seasonal ARIMA

Models
• Let us introduce several modifications made to the ARIMA model
to account for seasonal and nonstationary behavior.
• Often, the dependence on the past tends to occur most strongly at
multiples of some underlying seasonal lag S.
• For example, with monthly economic data, there is a strong yearly
component occurring at lags that are multiples of S= 12 because
of the strong connections of all activity to the calendar year.
• Data taken quarterly will exhibit the yearly repetitive period at s
= 4 quarters.
• Natural phenomena such as temperature also have strong
components corresponding to seasons.
• Hence, the natural variability of many physical, biological,
and economic processes tends to match with seasonal
fluctuations.
• Because of this , it is appropriate to introduce
autoregressive and moving average polynomials that
identify with the seasonal lags.
Pure Seasonal Autoregressive
seasonal
Moving Average modelautoregressi
ve operator
of order P
with
seasonal
• Where period seasonal
S.
moving
average
operator of
order Q with
seasonal
period S.
• Analogous to the properties of nonseasonal ARMA models,
• the pure seasonal
• is causal only when the roots of lie outside the unit circle,
• is invertible only when the roots of lie outside the unit
circle.
Illustration
• A first-order seasonal autoregressive moving average
series that might run
over months could be written as

or, exhibits the series in


terms of past lags at
the multiple of the
yearly seasonal
period s = 12 months
• estimation and forecasting for such a process involves
only straightforward modifications of the unit lag case
already treated.
• Causal if
• Invertible if |Θ|<1.
generalization
s of the
properties for
nonseasonal
models
First-Order Seasonal (s = 12)
MA model
• For the first-order seasonal (s = 12) MA model,

• it is easy to verify that


• Thus, the only nonzero correlation, aside from lag zero, is
• For the first-order seasonal (s = 12) AR model,
• we have

k=1,2,……
• In this case, the only non-zero correlations are
• ,k=0,1,2,….

• Again, the PACF have the analogous extensions from


nonseasonal to seasonal models.
• In general, we can combine the seasonal and nonseasonal
operators into
• a multiplicative seasonal autoregressive moving
average model, denoted by
• , and write

• as the overall model


Remarks
• In fitting such models, focusing on the seasonal
autoregressive and moving
• average components first generally leads to more
satisfactory results.
Illustration
• ARMA(0,1)
• Model:

• With | |<1 and |


Remark
• Seasonal nonstationarity can occur, for example, when the
process is nearly
• periodic in the season.
• For example, with average monthly temperatures over the
years, each January would be approximately the same,
• each February would be approximately the same, and so
on.
• In this case, we might think of average monthly
temperature as being modelled as

• where St is a seasonal component that varies slowly from


one year to the next, according to a random walk,
• The tendency of data to follow this type of model will be
exhibited in a sample ACF that is large and decays very
slowly at lags h = 12k,
• for k = 1,2,……
• If we subtract the effect of successive years from each
other, we find that

• This process is a stationary


• Its ACF will have a peak only at lag 12.
• In general, seasonal differencing can be indicated when
the ACF decays
• slowly at multiples of some season S, but is negligible
between the periods.
• Then a seasonal difference of order D is defined as

• where D = 1,2,3,…., takes positive integer values.


• Typically, D = 1 is sufficient to obtain seasonal
stationarity.
SARIMA
• The multiplicative seasonal autoregressive integrated
moving average model, or SARIMA model is given by

Usual
Gaussia
n white
noise
process.
• The general model is denoted as .
• The ordinary autoregressive and moving average
components are represented by polynomials and of orders
p and q, respectively
• The seasonal autoregressive and moving average
components by and of orders P and Q
• ordinary and seasonal difference components by
Example

• Model:
the multiplicative nature of the
model implies that the coefficient
of is the product of the
coefficients of and
rather than a free parameter.
• The multiplicative model assumption seems to work well
with many seasonal time series data sets
• while reducing the number of parameters that must be
estimated.

You might also like