ARIMA
ARIMA
Dr.K.D.Prasangika
Department of Mathematics
1
ARIMA (Box-Jenkins) Models
Introduction to ARIMA models
The Autoregressive Integrated Moving Average (ARIMA)
models, or Box-Jenkins methodology, are a class of linear
models that is capable of representing stationary as well as
nonstationary time series.
AR(1) AR(2)
ACF 0 ACF 0
PACF = 0 for lag > 1 PACF = 0 for lag > 2
Moving Average Models MA(q)
A qth-order moving average model, or MA(q), takes the form:
MA(1) MA(2)
ACF = 0 for lag > 1; PACF 0 ACF = 0 for lag > 2; PACF 0
ARMA(p,q) Models
A model with autoregressive terms can be combined with a model having
moving average terms to get an ARMA(p,q) model:
ACF PACF
Once a stationary series has been obtained, the analyst must identify the
form of the model to be used by comparing the sample ACF and PACF to
the theoretical ACF and PACF for the various ARIMA models.
ARIMA
280
270
260
The series show an
Index
250
upward trend.
240
230
220
210
1 6 12 18 24 30 36 42 48 54 60
Index
Autocorrelation Function for Index
(with 5% significance limits for the autocorrelations)
1,0
0,8
0,6
0,4
Autocorrelation
0,2
0,0
-0,2
-0,4
-0,6
-0,8
-1,0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Lag
ARIMA
Comparing the autocorrelations with their error limits, the only significant
autocorrelation is at lag 1. Similarly, only the lag 1 partial autocorrelation is
significant. The PACF appears to cut off after lag 1, indicating AR(1) behavior. The
ACF appears to cut off after lag 1, indicating MA(1) behavior we will try:
ARIMA(1,1,0) and ARIMA(0,1,1)
1,0 1,0
0,8 0,8
0,6 0,6
Partial Autocorrelation
0,4 0,4
Autocorrelation
0,2 0,2
0,0 0,0
-0,2 -0,2
-0,4 -0,4
-0,6 -0,6
-0,8 -0,8
-1,0 -1,0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Lag Lag
ARIMA
ARIMA(1,1,0) ARIMA(0,1,1)
1,0
0,8
0,6
0,4
Autocorrelation
0,2
0,0
-0,2
-0,4
-0,6
-0,8
-1,0
Autocorrelation Function for RESI2
(with 5% significance limits for the autocorrelations)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Lag 1,0
0,8
0,6
0,4
Autocorrelation
0,2
Finally, there is no significant residual 0,0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Therefore, either model is adequate Lag
70
60
The first sample ACF coefficient is significantly different form
50
zero. The autocorrelation at lag 2 is close to significant and
40
30
opposite in sign from the lag 1 autocorrelation. The remaining
20 autocorrelations are small. This suggests either an MA(1)
1 7 14 21 28 35
Index
42 49 56 63 70
model or an MA(2) model.
The first PACF coefficient is significantly different from zero,
but none of the other partial autocorrelations approaches
Autocorrelation Function for Readings
(with 5% significance limits for the autocorrelations) significance, This suggests an AR(1) or ARIMA(1,0,0)
1,0
0,8 Partial Autocorrelation Function for Readings
(with 5% significance limits for the partial autocorrelations)
0,6
0,4 1,0
Autocorrelation
0,2 0,8
0,0 0,6
Partial Autocorrelation
-0,2 0,4
-0,4 0,2
-0,6 0,0
-0,8 -0,2
-1,0 -0,4
-0,6
2 4 6 8 10 12 14 16 18
Lag -0,8
-1,0
2 4 6 8 10 12 14 16 18
Lag
ARIMA
A constant term is
MA(2) =
included in both
AR(1) = ARIMA(0,0,2)
ARIMA(1,0,0) models to allow for
the fact that the
readings vary about a
level other than zero.
1,0 1,0
0,8 0,8
0,6 0,6
0,4 0,4
Autocorrelation
Autocorrelation
0,2 0,2
0,0 0,0
-0,2 -0,2
-0,4 -0,4
-0,6 -0,6
-0,8 -0,8
-1,0 -1,0
2 4 6 8 10 12 14 16 18 2 4 6 8 10 12 14 16 18
Lag Lag
Therefore, either model is adequate and provide nearly the same three-step-
ahead forecasts. Since the AR(1) model has two parameters (including the
constant term) and the MA(2) model has three parameters, applying the
principle of parsimony we would use the simpler AR(1) model to forecast future
readings.