[Carlo Lazzaro]

Giorgio:
the usual fix is considering a more parsimonious model (ie, a lower number of predictors).

[Giorgio Di Stefano]

Thanks for the suggestions, Carlo. Indeed I do have a large number of predictors, which becomes larger due to the quadratic form and the factorials. I think will have to run a selection variables process.
Thank you for everything
Grazie di tutto!

[Carlo Lazzaro]

Giorgio:
if by selection variables process you mean -stepwise- (aka the contribute of the Lord of evil to statistics), please see Frank Harrel's warning list at https://www.stata.com/support/faqs/s...sion-problems/.
Conversely, if you mean skimming through the literature in your reserach field and see what others did in the past to give a fair and true view of the data generating process you're studying, it makes perfect sense.

[Giorgio Di Stefano]

Giorgio:
if by selection variables process you mean -stepwise- (aka the contribute of the Lord of evil to statistics), please see Frank Harrel's warning list at https://www.stata.com/support/faqs/s...sion-problems/.
Conversely, if you mean skimming through the literature in your reserach field and see what others did in the past to give a fair and true view of the data generating process you're studying, it makes perfect sense.

I mean the least absolute shrinkage and selection operator (lasso).
https://blog.stata.com/2019/09/09/an...asso-in-stata/

I have not used it before, nor have I a clue have to do it, but with more than a 100 predictors in total, I guess must figure out a way to reduce them
Does LASSO enter the same stepwise regression classification?

[Carlo Lazzaro]

Giorgio:
sorry, but I'm not familiar with -lasso-.

[Giorgio Di Stefano]

Giorgio:
sorry, but I'm not familiar with -lasso-.

Thank you for everything, Carlo!