1
$\begingroup$

I hope my question is not a dupplicate, I've searched on many forums and did not find the answer I'm looking for...

I'm currently studying Data Analysis, and I'm working on a project with time series.

But I'm struggling with the SARIMA Model, and how to set it up.

Here comes the questions : -Is a Dickey-Fuller test enough to know if a process is stationary ? -I was able to find the p, P, q, Q parameters with the autocorrelations functions, but I have no idea on how to find the d parameter ?

And finally, when I plot the autocorrelation function I get that :

Autocorrelation function : Autocorrelation function Partial autocorrelation function : Partial Autocorrelation Function Here is the Dickey-Fuller test result on my data :

ADF Statistic: -3.793415 p-value: 0.002975 Critical Values: 1%: -3.661 5%: -2.961 10%: -2.619

I'm guessing there's something wrong, but I did not find what... Any help will be more that appreciated !

Improve this question
$\endgroup$

1 Answer 1

Reset to default
0
$\begingroup$

ADF = Unit root test that cause stationarity, not because of seasonal effect.

One way to do would be to clear seasonality from your TS and then proceed ADF test.

But you have to keep in mind that H0 for ADF is presence of unit root, not stationarity.

You can have a ADF that rejects H0 while having a series non stationary because of seasonal effects.

Improve this answer
$\endgroup$
5
  • $\begingroup$ Hi, Thanks for your answer, that explains why my autocorrelation function looked this way. I removed the sasonality from my serie. I'm still struggling to know if it's stationary or not ; Is there any other test that I can run ? Thanks in advance $\endgroup$
    – Julien Paulet
    Commented Aug 7, 2019 at 16:26
  • $\begingroup$ Well apply the DF test on this new serie, if the pval is < 5% then you rejet presence of unit root, so your serie is stationary. $\endgroup$
    – josef_joestarr
    Commented Aug 7, 2019 at 17:46
  • $\begingroup$ post your data and I will try and help further $\endgroup$
    – IrishStat
    Commented Aug 8, 2019 at 10:25
  • $\begingroup$ Alright thanks a lot ! :) $\endgroup$
    – Julien Paulet
    Commented Aug 8, 2019 at 12:27
  • $\begingroup$ what happened to your data posting ? $\endgroup$
    – IrishStat
    Commented Oct 11, 2019 at 18:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.