Arima: Autoregressive Integrated Moving Average

ARIMA
Autoregressive Integrated
Moving Average
Introduction - ARMA
 ARMA - Auto Regressive Moving Average
 Introduced by Box and Jenkins in 1976.
 Box-Jenkins model.
 Used to develop a model that will forecast an
element based on its historical values.
 For example, the exchange rate in time t can
be forecasted based on its values in time t-2
and time t-5 plus stochastic error terms.
Two Common Processes
 Autoregressive process.
 Moving average process.
 It is likely that X has characteristics of both

AR and MA and is therefore ARMA. In
general, in an ARMA (p,q) process, there will
be p autoregressive and q moving average.
Introduction - ARIMA
 Most economic time series (i.e. GDP) are
nonstationary, that is, they are integrated.
 In general, if a time series has to be differenced d
times, it is integrated of order d or I(d).
 On the other hand, if d=0, the resulting to I(0)
process corresponds a stationary time series.
 Therefore, if we have to difference a time series d
times to make it stationary and then apply the
ARMA(p,q) model to it, we say that the original time
series is ARIMA (p,d,q).
P, D, Q Model
 P: the autoregressive parameter
 D: the integrated parameter of the number of
differencing passes
 Q: the moving average parameter
 For example: (0,1,2) model
The Constants in ARIMA
 If there are no autoregressive parameters in
the model, then the expected value of the
constant is , the mean of the series.
 If there are autoregressive parameters in the
series, then the constant represents the
intercept.
 If the series is differenced, then the constant
represents the mean or intercept of the
differenced series.
Assumptions
 Stationarity.
 No uncontrolled correlation.
 Arbitrary model lag order.
 No outliers.
 Randomly distributed shocks.
 Uncorrelated random errors.
Stationarity
 The input series should have a constant
mean, variance, and autocorrelation through
time.
 This assumption is tested through the
Augmented Dickey-Fuller Test and non-
stationarity is fixed through differencing.
Assumptions
No Uncontrolled Correlation
 Autocorrelation means that the value of a
given datum is largely determined by the
value of the preceding datum in the series.
 Assumption is tested through the Durbin-
Watson Coefficient, with range of value from
0 to 4. A value of 2 indicates no
autocorrelation, 0 indicates autocorrelation,
and 4 indicates negative autocorrelation.
Assumptions
Arbitrary Model Lag Order
 The researcher must have a theoretical basis
to establish the face validity of the order of
the model.
Assumptions
No Outliers
 As in other forms of regression, outliers may
affect conclusions strongly and misleadingly.
Assumptions
Randomly Distributed Shocks
 If shocks are present in the time series, they
are assumed to be randomly distributed with
a mean of 0 and a constant variance.
Assumptions
Uncorrelated Random Errors
 Residuals are randomly and normally
distributed, have non-significant
autocorrelations and partial autocorrelations,
and have a mean of 0 and homogeneity of
variance over time.
 The Durbin-Watson test is the standard test
for correlated error.
Assumptions
Procedure
 Test for the assumptions.
 The Box-Jenkins Methodology
 Identification.
 Estimation.
 Diagnostic Checking.
 Forecasting.
Test for Stationarity
 Visual plot.
 Correlogram.
 Unit root test.
 Augmented Dickey-Fuller Test
 Ho: Series is non-stationary
 Ha: Series is stationary.
 If absolute value of ADF > absolute value of the
critical regions, reject Ho
Differencing
 Differencing is a procedure which attempts to de-trend
the data in order to control autocorrelation and achieve
stationarity.
 It does this by subtracting each datum in a series from its
predecessor.
 The number of times a series needs to be differenced to
achieve stationarity is reflected in the d parameter.
 In order to determine the necessary level of differencing,
one should examine the plot of the data and
autocorrelogram
 Caution: Some time series may require little or no
differencing. An over differenced series produce less
stable coefficient estimates.
Procedure
Identification
 Major tools: ACF and PACF
 One autoregressive (p) parameter: ACF - exponential
decay; PACF - spike at lag 1, no correlation for other lags.
 Two autoregressive (p) parameters: ACF - a sine-wave
shape pattern or a set of exponential decays; PACF - spikes
at lags 1 and 2, no correlation for other lags.
 One moving average (q) parameter: ACF - spike at lag 1, no
correlation for other lags; PACF - damps out exponentially.
 Two moving average (q) parameters: ACF - spikes at lags 1
and 2, no correlation for other lags; PACF - a sine-wave
shape pattern or a set of exponential decays.
 One autoregressive (p) and one moving average (q)
parameter: ACF - exponential decay starting at lag 1; PACF
- exponential decay starting at lag 1.
Procedure
Estimation
 approximate maximum likelihood method
 the fastest method
 should be used for very long time series (e.g., with more
than 30,000 observations)
 approximate maximum likelihood method with
backcasting
 must use this method first to establish initial parameter
estimates that are very close to the actual final values
 exact maximum likelihood method
 may be inefficient when used to estimate parameters for
seasonal models with long seasonal lags (e.g., with yearly
lags of 365 days)
Procedure
Diagnostic Checking
 Test for the significance of the parameter
estimates.
 Use partial data to generate forecasts.
 Analysis of residuals.
Limitations
 The ARIMA method is appropriate only for a
time series that is stationary.
 At least 50 observations are recommended
for the input data.
 It is also assumed that the values of the
estimated parameters are constant
throughout the series.
Illustration
 Background: US GDP data from 1970 –
1996
 Frequency: quarterly
To Extract Data: File Open 
Foreign Data as Workfile Choose File
Open
Extract Data From Excel
1
2
3
Step 1: Check for Stationarity
To Plot Data: Series Box 
View Graph  Line
 Visual Plot
1
2
Step 1: Check for Stationarity
Series Box  View
 Unit Root Test  Level
 ADF / Unit Root Test
3
1
2
Step 2: Difference theSeries
Data Box  View
 Unit Root Test
 1st Difference
 ADF Test
3
2
Step 3: Estimate the P and Q
Series Box  View
Correlogram 
1st Difference
• Correlogram – ACF and PACF
3
2
Step 3: Estimate the P and Q
 Correlogram – ACF and PACF
Possible AR and MA models
And so on…
Quick Estimate Equation
Type equation: u_s_gdp c ar(1)
Step 4: Estimate
1 several
ok models
2
Quick Estimate Equation
Type Equation: u_s_gdp c ar(1) ma(1)
1  ok
2
Step 5: Determine the model
Model AIC value SC value
AR(1) 10.04537 10.10206m
AR(1)MA(1) 9.985905 10.07094
Step 6: Checking
 Use the t-test to check for the
significance of the parameters.
 Use the Durbin-Watson test to check for
the autocorrelation of the error terms.

Arima: Autoregressive Integrated Moving Average

Uploaded by

Copyright:

Available Formats

Arima: Autoregressive Integrated Moving Average

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Arima: Autoregressive Integrated Moving Average

Uploaded by

Copyright:

Available Formats

ARIMA

 Moving average process.

 It is likely that X has characteristics of both

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.