[R-package] Promote objective and init_score to top-level arguments in lightgbm()

[david-cortes]

ref #4968

This PR promotes objective and init_score to top-level parameters in lightgbm(). This is for 3 reasons:

• objective is probably one of the most common parameters to want to change. It's also accepted as a top-level parameter in the lgb.train interface and in the scikit-learn interface.
• Having objective as a top-level parameter will allow an option "auto" in the future (as part of [R-package] User-friendly redesign for lightgbm() #4968) to adjust it automatically for binary and multi-class classification depending on the input.
• Having init_score as a top-level parameter will allow to take it as a column from the input data in the future (as part of [R-package] User-friendly redesign for lightgbm() #4968), even if it right now looks out of place.

[jameslamb]

Thanks for this. I support adding init_score and objective as keyword arguments in lightgbm(). Please see the requested changes.

[david-cortes]

Updated.

[jameslamb]

Thanks very much for the tests! Please see my most recent requested changes.

[jameslamb]

Thanks very much! I think this is almost ready to merge, just a few more recommended changes.

[david-cortes]

Updated.

[jameslamb]

thanks! Just one more suggestion and this can be merged.

[jameslamb]

I just updated this with the latest changes from master, now that #5021 has been merged.

I'll merge this once it builds. Thanks very much for the contribution!

[github-actions]

This pull request has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this.