REG_ARIMA
Fits a univariate, non‑seasonal ARIMA time series model with the inclusion of one or more regression variables.
Required Arguments
Y — Array of length NOBS containing the time series. (Input)
IMODEL — Array of length 3 containing the model order parameters. (Input)
IMODEL(I) | Description |
|---|---|
1 | Order of the autoregressive part, p, where p≥0. |
2 | Order of the non‑seasonal difference operator, d, where d≥0. |
3 | Order of the moving average part, q, where q≥0. |
If p = 0 and q = 0, only regression is performed.
PARMA —Array of length 1 + p + q containing the estimated autoregressive (AR) and moving average (MA) parameters of the ARIMA(p,d,q) model. PARMA(1) is the estimated AR constant parameter, PARMA(2: (p+1)) contains the AR parameter estimates and PARMA((p+2):), contains the MA parameter estimates. (Output)
Optional Arguments
NOBS — Number of observations. (Input)
Default: NOBS = size(Y).
Default: NOBS = size(Y).
X — Array of size NOBS by K containing the regression data, where K = size(X,2) is the number of user supplied regression variables. (Input)
Specific columns in X may be selected using the INDX argument. Otherwise, all columns of X are used.
Default: No regression variables are included.
Specific columns in X may be selected using the INDX argument. Otherwise, all columns of X are used.
Default: No regression variables are included.
XLEAD — Array of size MXLEAD by K containing the regression data to be used in obtaining forecasts, where K = size(X,2) is the number of user supplied regression variables. (Input)
Specific columns in XLEAD may be selected using the INDX argument. Otherwise all columns of XLEAD are used.
Note: If MXLEAD > 0 and optional argument X is present, XLEAD is required.
Specific columns in XLEAD may be selected using the INDX argument. Otherwise all columns of XLEAD are used.
Note: If MXLEAD > 0 and optional argument X is present, XLEAD is required.
INDX — Index array containing the column numbers in X and XLEAD that are to be used for the regression variables. (Input)
Default: All columns of X and XLEAD are used.
Default: All columns of X and XLEAD are used.
TREND — Logical. If .TRUE., the routine will include a trend variable. (Input)
Note: Setting TREND = .TRUE. has the effect of fitting an intercept term in the regression. If the difference operator IMODEL(2) = d > 0, the effect on the model in the original, undifferenced space is polynomial of order d.
Default: TREND = .TRUE..
Note: Setting TREND = .TRUE. has the effect of fitting an intercept term in the regression. If the difference operator IMODEL(2) = d > 0, the effect on the model in the original, undifferenced space is polynomial of order d.
Default: TREND = .TRUE..
MXLEAD — Maximum lead time for forecasts. (Input)
Note: If MXLEAD > 1, forecasts are returned for t = NOBS + 1, .NOBS + 2, …, NOBS + MXLEAD
Default: MXLEAD = 0 (no forecasts).
Note: If MXLEAD > 1, forecasts are returned for t = NOBS + 1, .NOBS + 2, …, NOBS + MXLEAD
Default: MXLEAD = 0 (no forecasts).
MAXIT — Maximum number of iterations. (Input)
Default: MAXIT = 50.
Default: MAXIT = 50.
IPRINT — Printing option. (Input)
IPRINT | Action |
|---|---|
0 | No printing |
1 | Prints final results only |
2 | Prints intermediate and final results |
Default: IPRINT = 0.
PREG — Array of length K + t containing the estimated regression coefficients, where t=0 if TREND=.FALSE. or t=1 if TREND=.TRUE.. (Output)
SE — Array of length p + q containing the standard errors of the ARMA parameter estimates. (Output)
AVAR — White noise variance estimate. (Output)
Note: If IMODEL(0)+IMODEL(2)= 0 and K>0, AVAR is the mean squared regression residual.
Note: If IMODEL(0)+IMODEL(2)= 0 and K>0, AVAR is the mean squared regression residual.
REGSE — Array of length K + t containing the standard errors of the regression estimates, where t=0 if TREND=.FALSE. or t=1 if TREND=.TRUE.. (Output)
PREGVAR — Array of length (K + t) by (K + t) containing the variances and covariances of the regression coefficients, where t=0 if TREND=.FALSE. or t=1 if TREND=.TRUE.. (Output)
AIC — Akaike's Information Criterion for the fitted ARMA model. (Output)
LLIKE — Value of –2(ln(likelihood)) for fitted model. (Output)
FCST — Array of length MXLEAD containing the forecasts for time points t = NOBS + 1, .NOBS + 2, …, NOBS + MXLEAD. (Output)
FCSTVAR — Array of length MXLEAD containing the forecast variances for time points t = NOBS + 1, .NOBS + 2, …, NOBS + MXLEAD. (Output)
Fortran 90 Interface
Generic: CALL REG_ARIMA (Y, IMODEL, PARMA [, …])
Specific: The specific interface names are S_REG_ARIMA and D_REG_ARIMA.
Description
Routine REG_ARIMA fits an ARIMA(p, d, q) to a univariate time series with the possible inclusion of one or more regression variables.
Suppose Yt, t = 1, …, N, is a time series such that the d‑th difference is stationary. Further, suppose at is a series of uncorrelated, mean 0 random variables with variance .
The Auto‑Regressive Integrated Moving Average (ARIMA) model for {Yt, at} can be expressed as
where B is the backshift operator,
and
The notation for this model is ARIMA(p, d, q) where p is the order of the autoregressive polynomial , d is the order of the differencing needed to make stationary, and q is the order of the moving‑average polynomial .
The ARIMA model can be extended to include regression variables , by using the residuals (of the multiple regression of on ) in place of in the above ARIMA model.
Equivalently,
where
is the differenced residual series.
To estimate the (p + q + K) parameters of the specified regression ARIMA model, REG_ARIMA uses the iterative generalized least squares method (IGLS) as described in Otto, Bell, and Burman (1987).
The IGLS method iterates between two steps, one step to estimate the regression parameters via generalized least squares (GLS) and the second step to estimate the ARMA parameters. In particular, at iteration m, the first step finds
by solving the GLS problem with weight matrix
where
That is, minimizes , where 
, is an N by K matrix with i‑th column, , and is an N by N weight matrix defined using the theoretical autocovariances of the series . The series is modeled as an ARMA(p,q) with parameters and . At iteration m, the second step is then to obtain new estimates of and for the updated series,. To find the estimates and , REG_ARIMA uses the exact likelihood method as described in Akaike, Kitagawa, Arahata and Tada (1979) and used in routine, MAX_ARMA.
, is an N by K matrix with i‑th column, , and is an N by N weight matrix defined using the theoretical autocovariances of the series . The series is modeled as an ARMA(p,q) with parameters and . At iteration m, the second step is then to obtain new estimates of and for the updated series,. To find the estimates and , REG_ARIMA uses the exact likelihood method as described in Akaike, Kitagawa, Arahata and Tada (1979) and used in routine, MAX_ARMA. Comments
When forecasts are requested (MXLEAD > 0), REG_ARIMA requires that future values of the independent variables are provided in optional argument XLEAD. In effect, REG_ARIMA assumes the future X’s are known without error, which is valid for any deterministic function of time such as a seasonal indicator. Also, in economics, certain factors that are considered to be leading indicators are treated as deterministic for the purpose of predicting changes in the economy. Users may consider using a more general transfer function model if this is an unreasonable assumption. REG_ARIMA calculates forecast variances using the asymptotic result found in Fuller (1996) , Theorem 2.9.4. To obtain the standard errors of the ARMA parameters, REG_ARIMA calls routine NSLSE for the final w series.
Examples
Example 1
The data set consists of annual mileage per passenger vehicle and annual US population (in 1000’s) spanning the years 1980 to 2006 (U.S. Energy Information Administration, 2008). Consider modeling the annual mileage using US population as a regression variable.
use reg_arima_int
use umach_int
implicit none
integer, parameter :: mxlead=5, idep=2
integer :: i, nout, nobs, n
integer :: imodel(3)=(/1,0,0/), indind(1)=(/1/)
real(kind(1e0)) :: y(29), parma(2), xlead(mxlead, 2)
real(kind(1e0)) :: preg(2), regses(2)
real(kind(1e0)) :: ses(1), fcst(mxlead), fcstvar(mxlead)
real(kind(1e0)) :: avar,llike
! US mileage and population (1980-2008)
! Source: U.S. Energy Information
! Administration (Oct 2008).
real(kind(1e0)) :: x(29,2)
data x / &
22722.4681, 22946.5714, 23166.4458, 23379.1990, &
23582.4902, 23792.3795, 24013.2887, 24228.8918, &
24449.8982, 24681.923, 24962.2814, 25298.0941, 25651.4224,&
25991.8588, 26312.5820999999, 26627.8393, 26939.4284, &
27264.6925, 27585.4104, 27904.0168, 28217.1936, &
28503.9803, 28772.6647, 29021.0914, 29289.2127, &
29556.0549, 29836.2973, 30129.0332, 30405.9724, &
9062.0, 8813.0, 8873.0, 9050.0, 9118.0, 9248.0, 9419.0, &
9464.0, 9720.0, 9972.0, 10157.0, 10504.0, 10571.0, &
10857.0, 10804.0, 10992.0, 11203.0, 11330.0, 11581.0, &
11754.0, 11848.0, 11976.0, 11831.0, 12202.0, 12325.0, &
12460.0, 12510.0, 12485.0, 12293.0/
call umach(2,nout)
! Example 1
! The first column is the scaled US
! population and the second column is
! the annual mileage per vehicle
n = size(x,1)
nobs = n - mxlead
call scopy(nobs,x(1:n,idep),1,y,1)
call reg_arima(y, imodel ,parma, nobs=nobs, iprint=1, x=x, &
indx=indind, avar=avar, llike=llike, preg=preg, &
regse=regses, mxlead=mxlead, xlead=x(nobs+1:n,1:2), &
trend=.true., fcst=fcst, fcstvar=fcstvar, se=ses)
end
Output
Final results for regarima model (p,d,q) = 1 0 0
Final AR parameter estimates/ std errors
0.73063 0.13499
-2*ln(maximum log likelihood) = 231.8354
White noise variance 10982.5654
Regression estimates:
coef reg. se
1 -3481.22607 689.02661
2 0.54237 0.02673
Forecasts with standard deviation
t Y fcst Y fcst std dev
25 12360.40137 153.89659
26 12514.61035 171.89198
27 12673.53320 180.76634
28 12837.36719 185.32974
29 12991.26855 187.72037
Example 2
The data set consists of simulated weekly observations containing a strong annual seasonality. The seasonal variables are constructed and sent into REG_ARIMA as regression variables.
use reg_arima_int
use umach_int
use const_int
implicit none
integer, parameter :: mxlead = 4
integer :: nobs, i,n,idep,nout
integer :: imodel(3) =(/2,0,0/)
real(kind(1e0)) :: PI
real(kind(1e0)) :: preg(3),regses(3),parma(3)
real(kind(1e0)) :: ses(2),fcst(mxlead),fcstvar(mxlead)
real(kind(1e0)) :: avar,llike,aic,tmpr
real(kind(1e0)) :: x(100,2), xlead(mxlead,2)
real(kind(1e0)) :: y(104) =(/ &
32.27778,32.63300,33.13768,34.4517,34.63824, &
37.31262,37.35704,37.03092,36.39894,35.75541,&
35.10829,34.70107,34.69592,32.75326,30.85370,&
31.10936,29.47493,29.14361,28.50466,30.09714,&
28.49403,27.23268,23.49674,22.71225,21.42798,&
18.68601,17.40035,16.06832,15.31862,14.75179,&
13.40089,13.01101,12.44863,11.27890,11.51770,&
14.31982,14.67036,14.76331,15.35644,17.04353,&
18.39931,18.21919,18.72777,19.61794,22.31733,&
23.79600,25.41326,25.60497,27.93579,29.21765,&
29.60981,28.46994,28.78081,30.96402,35.49537,&
35.75124,36.18933,37.2627,35.02454,33.57089, &
35.00683,34.83886,34.19827,33.73966,34.49709,&
34.07127,32.74709,31.97856,31.3029,30.21916, &
27.46015,26.78431,25.32815,23.97863,21.83837,&
21.00647,20.58846,19.94578,17.38271,17.12572,&
16.71847,17.45425,16.15050,13.07448,12.54188,&
12.42137,13.51771,14.84232,14.28870,13.39561,&
15.48938,16.47175,17.62758,16.57677,18.20737,&
20.8491,20.15616,20.93857,23.73973,25.30449, &
26.51106,29.43261,32.02672,32.18846/)
call umach(2,nout)
PI = const('PI')
! The data are simulated weekly
! observations with an annual
! seasonal cycle
n = size(y)
nobs = n - mxlead
! Create the seasonal variables
do i = 1, nobs
x(i,1)=sin(2*PI*i/float(52))
x(i,2) = cos(2*PI*i/float(52))
end do
do i=1, mxlead
xlead(i,1)=sin(2*PI*(i+nobs)/float(52))
xlead(i,2) = cos(2*PI*(i+nobs)/float(52))
end do
call reg_arima(y, imodel, parma, iprint=1, x=x, &
nobs=nobs, avar=avar, llike=llike, &
preg=preg, regse=regses, mxlead=mxlead, &
xlead=xlead, trend=.true., fcst=fcst, &
fcstvar=fcstvar, se=ses)
end
Output
Final results for regarima model (p,d,q) = 2 0 0
Final AR parameter estimates/ std errors
0.71860 0.09836
-0.25991 0.09827
-2*ln(maximum log likelihood) = -13.6209
White noise variance 0.8783
Regression estimates:
coef reg. se
1 24.81010 0.17175
2 9.68013 0.23993
3 5.72305 0.24751
Forecasts with standard deviation
t Y fcst Y fcst std dev
101 26.74492 1.31358
102 28.07805 1.47616
103 29.33707 1.49560
104 30.53160 1.49560
