0
$\begingroup$

I have time series data which is measured in each 5 second. I am trying to forecast the data. But, ACF and PACF results are seems to different than other plot that I have seen. I am using statsmodels as a library like below.

from statsmodels.graphics.tsaplots import plot_pacf
plot_pacf(ts, lags = 40, method = "ols");

from statsmodels.graphics.tsaplots import plot_acf
plot_acf(ts, lags = 50);

enter image description here

enter image description here

What I need to interpret with these plots?

Improve this question
$\endgroup$
4
  • $\begingroup$ Your series might have a unit root. What does sn stand for? $\endgroup$
    – Richard Hardy
    Commented Feb 22, 2023 at 17:58
  • $\begingroup$ Hi Richard, I fixed it. Measurements happens in each 5 second. $\endgroup$
    – MrsHelios
    Commented Feb 23, 2023 at 13:32
  • $\begingroup$ With the ADF test, I can said that data is non-stationary. But, Is the PACF, ACF reasonable for the unit root case? $\endgroup$
    – MrsHelios
    Commented Feb 23, 2023 at 13:38
  • 1
    $\begingroup$ Yes, a unit root process may produce ACF and PACF that look like yours. $\endgroup$
    – Richard Hardy
    Commented Feb 23, 2023 at 14:40

0

Reset to default

Your Answer

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

Browse other questions tagged or ask your own question.