Thread: help about ARMA model

Hi guys !
This is my first post and i have a sort of a dumb question :

Do the time serie need to be stationary in order to model it to ARMA ??

actually i'm trying to modelise "DEATHS.tsm" with ITSM and i have calculated \nabla\nabla^{12}X_{t}

but when i do the tests of randomness, and exactly the ljung-box test the p-value is : 0.011

should i continue and modelise it ???


Thnx in advance

Hey Nooob.

Does the randomness reflect the independence of the independence of the error terms? If it does then the data do not reflect the assumptions of the model.

so i can modelize it ??

What is the definition of your hypothesis in word and algebraic form?

I opened "DEATHS.tsm" on ITSM
I have detrended and desesionalized it
you can download the modified serie from here :deaths.tsm
my problem is : i don't know if i can start to modelize the serie or should i make more modifications to the serie ??

Can you answer the above question please?

i can't put the problem in an algebreca form
cause my problem is in the software , i don't know how to make you understand
plz read my last question , i have explained the problem in there

and i'm following the method of box-junkis to modelise this serie

wating for help

I can't help you if you don't provide the information asked for above.