9:15 p.m.
Hello,
I tested your script and data, without the --x-12-arima option (because I'm
with Linux), and worked fine:
gretl versão 1.10.0cvs
Sessão atual: 2014-09-30 22:11
? open "Exemplo.gdt"
Lido o ficheiro de dados /home/helio/gretl/Exemplo.gdt
periodicidade: 12, máxobs: 480
intervalo das observações: 1980:01 a 2019:12
Lista das 2 variáveis:
0) const 1) ipc_dive_m
? smpl 2004:10 2011:10
Intervalo completo dos dados: 1980:01 - 2019:12 (n = 480)
Amostra atual: 2004:10 - 2011:10 (n = 85)
? arima 0 0 1 ; 1 0 0 ; ipc_dive_m
Funções calculadas: 46
Cálculos de gradientes: 12
Modelo 1: ARMA, usando as observações 2004:10-2011:10 (T = 85)
Estimado usando o filtro de Kalman (Máxima Verosimilhança (ML) exacta )
Variável dependente: ipc_dive_m
Erros padrão baseados na Hessiana
coeficiente erro padrão z valor p
---------------------------------------------------------
const 0,407476 0,0742870 5,485 4,13e-08 ***
Phi_1 −0,0507680 0,108372 −0,4685 0,6395
theta_1 0,549245 0,0886602 6,195 5,83e-10 ***
Média var. dependente 0,410204 D.P. var. dependente 0,544479
Média de inovações 0,001229 D.P. das inovações 0,463257
Log. da verosimilhança −55,39961 Critério de Akaike 118,7992
Critério de Schwarz 128,5698 Critério Hannan-Quinn 122,7292
Real Imaginária Módulo Frequência
-----------------------------------------------------------
AR (sazonal)
Raiz 1 -19,6974 0,0000 19,6974 0,5000
MA
Raiz 1 -1,8207 0,0000 1,8207 0,5000
-----------------------------------------------------------
# --x-12-arima
? fcast 2011:11 2011:11
Para intervalos de confiança a 95%, z(0,025) = 1,96
ipc_dive_m predição erro padrão intervalo a 95%
2011:11 0,23 0,36 0,463 -0,55 - 1,27
Estatísticas da avaliação da Predição
Erro Médio -0,13221
Erro Médio Quadrado 0,017479
Erro Unitário Médio Quadrado 0,13221
Erro Médio Absoluto 0,13221
Erro Médio Percentual -57,482
Erro Médio Percentual Absoluto 57,482
U de Theil 0
---
Hélio
On Tue, Sep 30, 2014 at 9:29 PM, <henrique.andrade(a)bb.com.br> wrote:
Dear Gretl Team,
Please take a look at the following (very simple) script:
<hansl>
open "Exemplo.gdt"
smpl 2004:10 2011:10
arima 0 0 1 ; 1 0 0 ; ipc_dive_m --x-12-arima
fcast 2011:11 2011:11
<hansl>
It gives me infinite values for the sum of squared residuals and a very
strange behaviour of the forecasted value (9,87378e+307). The ipc_dive_m is
an inflation measure from Brazil.
Best regards,
*Henrique Coêlho de Andrade*
Diretoria de Estratégia e Organização
Divisão de Cenários e Estudos Macroeconômicos
Banco do Brasil
henrique.andrade(a)bb.com.br
Tel.: (61) 3102-6911
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
9:51 p.m.
On Tue, 30 Sep 2014, henrique.andrade(a)bb.com.br wrote:
Dear Gretl Team,
Please take a look at the following (very simple) script:
<hansl>
open "Exemplo.gdt"
smpl 2004:10 2011:10
arima 0 0 1 ; 1 0 0 ; ipc_dive_m --x-12-arima
fcast 2011:11 2011:11
<hansl>
It gives me infinite values for the sum of squared residuals and a very
strange behaviour of the forecasted value (9,87378e+307). The ipc_dive_m is
an inflation measure from Brazil.
Confirmed here. The strange numbers are obviously the result of an
NA value getting involved in the calculations where this was not
expected and therefore not handled properly. The next question is
how/why did the NA value arise?
As Hélio has remarked, this doesn't happen when the model is
estimated via gretl's native arima functionality. So it's something
to do with our interface with X-12-ARIMA (or X13-ARIMA-SEATS). I'm
working on it.
Allin Cottrell
3:41 p.m.
On Tue, 30 Sep 2014, henrique.andrade(a)bb.com.br wrote:
Please take a look at the following (very simple) script:
<hansl>
open "Exemplo.gdt"
smpl 2004:10 2011:10
arima 0 0 1 ; 1 0 0 ; ipc_dive_m --x-12-arima
fcast 2011:11 2011:11
<hansl>
It gives me infinite values for the sum of squared residuals and a very
strange behaviour of the forecasted value [...]
This is now fixed in CVS and snapshots.
Allin Cottrell
3495
days inactive
3497
days old
4 comments
3 participants
participants (3)
-
Allin Cottrell -
henrique.andrade＠bb.com.br -
Hélio Guilherme