add syntactic cases for xgboost

[razdoburdin]

Description

This PR adds few syntactic cases for xgboost regression benchmarks

Checklist:

Completeness and readability

• Git commit message contains an appropriate signed-off-by string (see CONTRIBUTING.md for details).
• I have resolved any merge conflicts that might occur with the base branch.

Testing

• I have run it locally and tested the changes extensively.
• All CI jobs are green or I have provided justification why they aren't.

[david-cortes-intel]

Wasn't this meant to replace a sparse dataset?

Also, in order to make it more realistic, how about adding more noise, irrelevant variables, skewed distributions, and so on? The first two are controllable as arguments to make_regression, while the latter would require manually applying transformations (e.g. binning).

[razdoburdin]

Wasn't this meant to replace a sparse dataset?

Also, in order to make it more realistic, how about adding more noise, irrelevant variables, skewed distributions, and so on? The first two are controllable as arguments to make_regression, while the latter would require manually applying transformations (e.g. binning).

This is the case, I have used for EMR vs GNR perf comparison. We can add more realistic cases latter.

[david-cortes-intel]

Do these parameters make sense without reducing the learning rate? I guess in this case it'd be a toy problem with very high predicatibility, but would the tree structure end up being similar as what you'd get for the original epsilon data?

[razdoburdin]

sorry for misunderstanding.

this is not the replacement for epsion. This is the toy case I have used for EMR vs GNR performance comparison. We need to have this case in some public benchmarks to be able to share it.