Differencing time series outside TS ARIMA

[ritikat]

Hi Experts,

Can we do differencing on a time series outside TS ARIMA tool? I want to difference the series outside and then check the ACF and PCF in the TS PLOT. 

Regards,

Ritika

[AngelosPachis]

Hi @ritikat ,

Not sure if that answers the question, but can't you imitate differencing with a multi-row formula tool  (first, second grade if needed and so on) and then checking if the time-series is stationary or not based on that?

It's a really interesting question. If I may ask, what's the reason that makes you want to do it outside the ARIMA tool?

Regards,

Angelos

[ritikat]

Hi @AngelosPachis ,

Thank you for your prompt answer

The idea is to analyze the following in the time series - 

1) If the time series is stationary or not - I did a Dicky Fuller's test using python. After checking the ADF coefficient and p - value , I figured that series is not stationary. 

2) Make the time series stationary and then again do the ADF test to check if it's stationary. To do this step, I would like to do the differencing outside. 

Regards,

Ritika

[AngelosPachis]

Oh I see, to be honest I am not sure how the ARIMA tool enforces/checks stationarity and whether it uses the Dickey Fuller test. I know it is somehow done automatically but can't quite tell.

I think the multi-row formula tool will help you do the differencing outside of the tool and then feed the data into Python. Hope this will work, it's a really cool idea