Professional Documents
Culture Documents
Umberto Triacca
Dipartimento di Ingegneria e Scienze dellInformazione e Matematica
Universita dellAquila,
umberto.triacca@ec.univaq.it
'$
DGP x (k)
&%
7
?
x1 , ..., xT
Umberto Triacca Lesson 9: Autoregressive-Moving Average (ARMA) models
Introduction
xt ARMA(p, q),
if
(z) = 1 1 z ... p z p
and
(z) = 1 + 1 z... + q z q
have no common factors.
Umberto Triacca Lesson 9: Autoregressive-Moving Average (ARMA) models
Autoregressive-Moving Average (ARMA) models
xt = 0.3xt1 + ut + 0.7ut1 t Z,
xt = 0.7xt1 0.5xt1 + ut t Z,
xt = ut + 0.7ut1 t Z,
where ut WN(0, u2 ).
and
(z) = 1 0.7z
We note that both polynomials have a common factor, namely
1 0.7z. Discarding the common factor in each leaves
(z) = 1 0.3z
and
(z) = 1.
Thus the process is an AR(1) process, defined by
xt = 0.3xt1 + ut
Umberto Triacca Lesson 9: Autoregressive-Moving Average (ARMA) models
Causal Autoregressive-Moving Average (ARMA)
models
and
X
xt = j utj t.
j=0
It means that
!2
n
X
lim E xt j utj = 0.
n
j=0
xt = 3xt1 + ut
xt = ut + ut1
and
X
ut = j xtj t.
j=0
1 , 2 , ..., p , 1 , 2 , ..., q , u2
'$