FAQ: Can we use diagnostics and regression plots after other models than OLS?

[josef-pkt]

not yet, or at least not verified.

Post-estimation results like residual diagnostics, regression plots and outlier-influence measures have been written for plain OLS with some new additions to linear time series models.

Some of those are applicable to any type of residuals, but in many cases some parts are specific to OLS. However, even if the functions are applicable for other residuals or models, those results have not been verified and checking whether the functions are appropriate is still up to the user.

e.g. regression plots don't seem to work with penalized estimation even using penalized OLS
http://stats.stackexchange.com/questions/252733/partial-regression-plots-for-regularized-l2-linear-regression?noredirect=1#comment481531_252733