Hi there,
merely to straighten out a wee point first of all, I ask the following by way of trying to help.
Are you aware (for example) that in their books, P Brockwell and R Shumway pose the
ARIMA equations in different form re the positive and negative coefficient signage.
Their models are different and the signage results are different but
when the coefficients are inserted in their models, the results are the same.
Just asking that you properly know the model being computed is all I'm asking. 
 
rjfhud@powerup.com.au
----- Original Message -----
From: 不提供 不提供
To: gretl-users@lists.wfu.edu
Sent: Sunday, January 09, 2011 10:54 PM
Subject: [Gretl-users] Some questions about X-12-ARIMA


Dear all:

I make my question clearer. ARIMA and X-12-ARIMA have almost the same outcomes under most combinations of AR and MA. For example, Using the same sample, the output of ARIMA(1,1,1)(1,1,0 ):
 
Function evaluations: 22
Evaluations of gradient: 8
Model 5: ARIMA, using observations 1982:03-1989:12 (T = 94)
Estimated using BHHH method (conditional ML)
Dependent variable: (1-L)(1-Ls) z
 
             coefficient   std. error      z       p-value
  ---------------------------------------------------------
  phi_1       0.0386387     0.490287     0.07881   0.9372 
  Phi_1      -0.547450      0.103980    -5.265     1.40e-07 ***
  theta_1     0.134454      0.505469     0.2660    0.7902 
Mean dependent var  -595.9894   S.D. dependent var   35113.05
Mean of innovations -657.4065   S.D. of innovations  29171.20
Log-likelihood      -1099.788   Akaike criterion     2207.577
Schwarz criterion    2217.750   Hannan-Quinn         2211.686
                        Real  Imaginary    Modulus  Frequency
  -----------------------------------------------------------
  AR
    Root  1          25.8808     0.0000    25.8808     0.0000
  AR (seasonal)
    Root  1          -1.8266     0.0000     1.8266     0.5000
  MA
    Root  1          -7.4375     0.0000     7.4375     0.5000
  -----------------------------------------------------------
 
the output of X-12-ARIMA(1,1,1)(1,1,0 ):
 
Model 6: ARIMA, using observations 1982:03-1989:12 (T = 94)
Estimated using X-12-ARIMA (conditional ML)
Dependent variable: (1-L)(1-Ls) z
             coefficient   std. error      z       p-value
  ---------------------------------------------------------
  phi_1       0.0383739    0.602274      0.06371   0.9492 
  Phi_1      -0.547423     0.0911210    -6.008     1.88e-09 ***
  theta_1     0.134554     0.597619      0.2252    0.8219 
Mean dependent var  -595.9894   S.D. dependent var   35113.05
Mean of innovations -657.4774   S.D. of innovations  29171.20
Log-likelihood      -1099.788   Akaike criterion     2207.577
Schwarz criterion    2217.750   Hannan-Quinn         2211.686
                        Real  Imaginary    Modulus  Frequency
  -----------------------------------------------------------
  AR
    Root  1          26.0594     0.0000    26.0594     0.0000
  AR (seasonal)
    Root  1          -1.8267     0.0000     1.8267     0.5000
  MA
    Root  1          -7.4320     0.0000     7.4320     0.5000
  -----------------------------------------------------------
 
 
The outcomes of ARIMA(1,1,1)(1,1,0 ) and X-12-ARIMA(1,1,1)(1,1,0 ) are almost the same.

But there are a few exceptions. For example, under the same sample, the output of ARIMA(1,1,2)(2,1,0 ):
 
Model 7: ARIMA, using observations 1983:03-1989:12 (T = 82)
Estimated using BHHH method (conditional ML)
Dependent variable: (1-L)(1-Ls) z
             coefficient   std. error     z      p-value
  -------------------------------------------------------
  phi_1       -0.590308    0.200862     -2.939   0.0033   ***
  Phi_1       -0.683313    0.134247     -5.090   3.58e-07 ***
  Phi_2       -0.240713    0.113586     -2.119   0.0341   **
  theta_1      0.873512    0.207170      4.216   2.48e-05 ***
  theta_2      0.361254    0.0966288     3.739   0.0002   ***
Mean dependent var  -1074.305   S.D. dependent var   36698.54
Mean of innovations -1019.087   S.D. of innovations  28580.42
Log-likelihood      -957.7121   Akaike criterion     1927.424
Schwarz criterion    1941.864   Hannan-Quinn         1933.222
                        Real  Imaginary    Modulus  Frequency
  -----------------------------------------------------------
  AR
    Root  1          -1.6940     0.0000     1.6940     0.5000
  AR (seasonal)
    Root  1          -1.4194    -1.4628     2.0382    -0.3726
    Root  2          -1.4194     1.4628     2.0382     0.3726
  MA
    Root  1          -1.2090    -1.1430     1.6638    -0.3795
    Root  2          -1.2090     1.1430     1.6638     0.3795
  -----------------------------------------------------------
 
the output of X-12-ARIMA(1,1,2)(2,1,0 ):
 
Model 8: ARIMA, using observations 1983:03-1989:12 (T = 82)
Estimated using X-12-ARIMA (conditional ML)
Dependent variable: (1-L)(1-Ls) z
             coefficient   std. error     z      p-value
  -------------------------------------------------------
  phi_1        0.653709     0.209156     3.125   0.0018   ***
  Phi_1       -0.675406     0.113095    -5.972   2.34e-09 ***
  Phi_2       -0.244173     0.113191    -2.157   0.0310   **
  theta_1     -0.566737     0.220105    -2.575   0.0100   **
  theta_2     -0.222901     0.115118    -1.936   0.0528   *
Mean dependent var  -1074.305   S.D. dependent var   36698.54
Mean of innovations -2724.431   S.D. of innovations  29295.00
Log-likelihood      -959.7371   Akaike criterion     1931.474
Schwarz criterion    1945.914   Hannan-Quinn         1937.272
                        Real  Imaginary    Modulus  Frequency
  -----------------------------------------------------------
  AR
    Root  1           1.5297     0.0000     1.5297     0.0000
  AR (seasonal)
    Root  1          -1.3830     1.4774     2.0237     0.3698
    Root  2          -1.3830    -1.4774     2.0237    -0.3698
  MA
    Root  1           1.1990     0.0000     1.1990     0.0000
    Root  2          -3.7416     0.0000     3.7416     0.5000
  -----------------------------------------------------------
 
 
The outcomes of ARIMA(1,1,2)(2,1,0 ) and X-12-ARIMA(1,1,2)(2,1,0 ) are hugely different.
The question above puzzles me.

I also want to know When I choose the options Model/Time series/ARIMA/Using X-12-ARIMA to run the X-12-ARIMA model. Is the set of equation of X-12-ARIMA in gretl the same as RegARIMA(X-12-ARIMA – Reference Manual, Version 0.3. U.S. Census Bureau):
φ(B)Φ(B)▽^d ▽_s^D[y-Σβ_i x_it]= θ(B)Θ(B)a_t
I can not see the outcome of any seasonality adjusting regression variables(the part of y-Σβ_i x_it, such as length-of-month、Trend constant、Trading day、level shift at t_0 and so on).
 
Thanks a lot     

_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users