Академический Документы
Профессиональный Документы
Культура Документы
Introduction
If random variable X is indexed to time, usually denoted by t,
the observationsX t , t
Stationarity:
Strictly /strongly stationarity
Weak/ loose (wide-sense) stationarity
Wide-sense Stationarity (Secondary stationarity)
Definition 2: The time seriesX t , t ( is the integer set) is
said to be stationary if:
(i) E X 2t ; t
(ii) E X t ; t
(iii) X t,s X t h,s h ; t,s, h
k t, t k Cov X t , X t k E X t t X t k t k
For stationary process {Xt}, then we have the mean
E(Xt)=E(Xt+k)=μ and variance Var(Xt)=σ2. Hence, the
autocovariance:
k t, t k Cov X t , X t k E X t X t k
Autocorrelation Function
The autocorrelation function (ACF) is
k
k Corr X t , X t k
0
for integers k, 0 Cov X t , X t k Var X t
so that:
E X t X t k
k
E Xt
2
PROPERTIES THE AUTOCOVARIANCE AND THE
AUTOCORRELATION FUNCTIONS
PROPERTIES:
1. 0 Var X t 0 1.
2. k 0 k 1.
3. k k and k k , k.
4. (necessary condition) k and k are positive semi-
definite
n n
i j ti t j 0
i 1 j 1
n n
i j ti t j 0
i 1 j 1
for any set of time points t1,t2,…,tn and any real numbers 1,2,…,n.
10
THE PARTIAL AUTOCORRELATION FUNCTION (PACF)
X t k k1X t k 1 k 2 X t k 2 kk X t et k
12
CALCULATION OF PACF
and taking the expectations
PACF
13
CALCULATION OF PACF
For j=1,2,…,k, we have the following system of equations
1 k1 k 2 1 kk k 1
2 k11 k 2 kk k 2
k k1 k 1 k 2 k 2 kk
14
CALCULATION OF PACF
Using Cramer’s rule successively for k=1,2,…
11 1
1 1
1 2 2 12
22
1 1 1 12
1 1
15
CALCULATION OF PACF
1 1 2 k 2 1
1 1 1 k 3 2
k 1 k 2 k 3 1 k
kk
1 1 2 k 2 k 1
1 1 1 k 3 k 2
k 1 k 2 k 3 1 1
16
CALCULATION OF PACF
2. Levinson and Durbin’s Recursive Formula:
k 1
k k 1, j k j
j 1
kk k 1
1 k 1, j k j
j 1
17
Reference
STAT 497_LN2
G.P. Nanson, Stationary and non-stationary time series
(chapter 26).