[R] Problems with ARIMA commands

[Livio Fenga]

Hi!
I have to fit 36 ARMA models  (all combination from the order (0,0) to 
the order (5,5) ) to 1000 bootstrap replication from each of 1000 
simulated time series following an ARMA(2,1) process. The simulated 
series are generated by ARIMA.SIM command.

The problem is the following: the esimation procedure of ARMA 
coefficient stops after the estimation of a certain number of bootstrap 
replication (it ranges from 10,15 to 55) always before the 10-th 
simulated serie due to an "error in solve.default(res$hessian * n.used) 
:  Lapack routine dgesv: is   the system is exactely singiular" .

I tried to modify the command OPTIM in the command ARIMA, it worked but 
not when  I try to apply this modified OPTIM in the ARMA command. In 
other words I put an error in the OPTIM command but ARMA ignores it and 
gives me parameters estimation.

Another (related) question is the following:  is there a way to modify 
ARIMA command in order to get only the estimated coefficients and the AIC ?

I really would like to thank for the kind attention.

Best,

Livio Fenga



Chiacchiera con i tuoi amici in tempo reale! 
 http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com