# Estimate a regression model with ARMA errors?

**URL:** <https://community.wolfram.com/t/estimate-a-regression-model-with-arma-errors/16050>\
**Category:** Questions\
**Tags:** data-science, economics, finance\
**Created:** [May 22, 2020, 5:16am UTC](https://community.wolfram.com/t/estimate-a-regression-model-with-arma-errors/16050 "2020-05-22T05:16:53Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![qunyongwang](https://avatars.discourse-cdn.com/v4/letter/q/d2c977/32.png) [@qunyongwang](https://community.wolfram.com/u/qunyongwang)\
**Post date:** [May 22, 2020, 5:16am UTC](https://community.wolfram.com/t/estimate-a-regression-model-with-arma-errors/16050/1 "2020-05-22T05:16:53Z")

</div>

Mathematica provides a good estimation method for pure time series process and regression model. But how to estimate a regression model with ARIMA errors? The model is

y = x\*b + u,  
where u ~ ARIMA process.

I think it is big shortage for Mathematica to estimate some regression model. I believe it can be improved a lot if it allows the users to specify some exogenous variables in its existing distribution functions.

Thanks.