ARMA checking
Posted: Sat Sep 12, 2009 12:36 pm
Hi
i am a new member and I have a question pertaining to ARMA model. I can not completely understand what I need to do. I know that first is identification- to pick up the appropriate orders p and q.I construct collegrame and saw that all ac and pac coefficient were outside of the confidence interval. How can I interpret this result. Do I need to make a conclusion that I can not use mixed ARMA? Using intimation criteria( SBIC) I pick up model ARMA(0,0). What is the interpretation of this model when you have p=0, q=0.Finally to check whether my model is fitted I know I need to use Box- Ljung test. I do not know how exactly I can run this test only for the residuals. And then if I see that hypothesis that autorelation coefficient are equal to zero can not be rejected what should I confirm with this diagnostic test.
i am a new member and I have a question pertaining to ARMA model. I can not completely understand what I need to do. I know that first is identification- to pick up the appropriate orders p and q.I construct collegrame and saw that all ac and pac coefficient were outside of the confidence interval. How can I interpret this result. Do I need to make a conclusion that I can not use mixed ARMA? Using intimation criteria( SBIC) I pick up model ARMA(0,0). What is the interpretation of this model when you have p=0, q=0.Finally to check whether my model is fitted I know I need to use Box- Ljung test. I do not know how exactly I can run this test only for the residuals. And then if I see that hypothesis that autorelation coefficient are equal to zero can not be rejected what should I confirm with this diagnostic test.