Create error handling function in linear regression algorithm of ML

[pahalsrivastava]

What would you like to share?

We can add error handling function to the linear regressio algorithm in the ML folder; so that it can handle errors and still run the linear regression algorithm

Additional information

No response

[tianyizheng02]

The linear regression algorithm in this repo should be reimplemented anyway (see #8847). As such, this additional handling should be either incorporated into the reimplementation PR or saved for a later PR.

In any case, what sort of error handling are you suggesting? Input validation, floating-point errors, etc?