confusion in calculation of e_t terms in ARMAX model

9 views

Skip to first unread message

jay katkani

unread,

Jul 10, 2022, 3:32:47 PM7/10/22

to pystatsmodels

I have generated a SARIAMX(0,0,2) model and forecasted the model on in-sample data using python statsmodels library.

I want to cross-validate the model.predict() result using ARMAX equation, for that i used the coefficients from model.summary() and used that equation for in-sample forecasting but both results( results from python and result generated manually in excel) are not matching

I tried surfing the net figuring out what i am doing wrong but couldn't figure it out.

I am using meantemp as dependent variable and humidity and windspeed as exogenous variable.
I am attaching my google sheet link where i did my calculation and i am also attaching the model.summary() in the sheet itself.
link of sheet : https://docs.google.com/spreadsheets/d/1yRicuVWl9P0QQRG-PMKGthu0BowhFcsAAKSeAvs0cw0/edit?usp=sharing