The moving average process of order , MA(), is defined as
|
(12.1) |
with white noise
. With the Lag-Operator (see
Definition 10.13) instead of (11.1) we can write
|
(12.2) |
with
The MA()
process is stationary, since it is formed as the linear
combination of a stationary process. The mean function is simply
. Let , then the covariance structure is
For the ACF we have for
|
(12.3) |
and
for
, i.e., the ACF breaks off after
lags.
As an example consider the MA(1) process
which according to (11.3) holds that
and
for . Figure
11.1 shows the correleogram of a MA(1) process.
Fig.:
ACF of a MA(1) process with (left)
and
(right).
SFEacfma1.xpl
|
Obviously the process
has the same ACF, and it holds that
In other words the process with the parameter has the same
stochastic properties as the process with the parameter .
This identification problem can be countered by requiring that the
solutions of the characteristic equation
|
(12.4) |
lie outside of the complex unit circle. In this case the linear
filter is invertible, i.e., there exists a polynomial
so that
and
Figure
11.2 displays the correlogram of a MA(2) process
for some collections of parameters.
Fig.:
ACF of a MA(2) process with
(top left),
(top right),
(bottom left) and
(bottom right).
SFEacfma2.xpl
|