Add cross validation to LGBMTuner

[mirix]

Some sort of LGBMTunerCV being able to customise the number of CVs would be great.

[DanilZherebtsov]

Yeah, this is the second request on this matter in the past few months. Working on implementing the sklearn cv modules into LGBMTuner pipeline.

One thing to have in mind - if you don't have a specific need for custom validation strategy and just want to make you random tuning validation more robust - LGBMTuner has this covered as is: every trial of hyperparameters optimisation is carried out on a new random split. This way this is very similar to a cross-validation approach you are seeking. If you are using LGBMTuner with default parameters (200 trials) - that means you will have 200 random train/valid splits during the tuning process.

Does that answers your question, or you have specific reasons for custom validation strategy?