time series stationarity

esraa1 — Thu, 02 Jul 2020 20:24:27 GMT

Hi!

from ACF and PACF plots, as shown below , the data is stationary or not ? why ?

Re: time series stationarity

gamotte — Fri, 03 Jul 2020 10:34:26 GMT

Hello,

Shouldn't the PACF be 1 at 0 ?

I would say that the data is stationary as if wasn't the ACF (Edit: or PACF) would be decreasing more slowly.

Most peaks are not significant and there seems to be some seasonal effect.

Maybe you can try to fit an ARIMA (1,0,0)x(1,0,0)12 and see if the residuals are white noise (no significant peaks).

Those are just my blurry memories from university many years ago. You will probably have more valid answers by posting your question

on the Forecasting and Econometrics forum :

Edit: For some reason, the link doesn't work. You can find the forum in the Analytics menu.

esraa1 — Fri, 03 Jul 2020 10:47:08 GMT

thanks,

the table below is the result of dickey fuller test, show that the data not stationary at lag 2.