optimize_gbm_parameters: Optimizing GBM parameters.
In Bilot/AI-jack-opensource-R: Package for Machine Learning project handling

Optimazation is done in two steps: (1) maximum depth is optimized, (2) the rest of the parameters are optimized. This procedure makes the optimization problem more manageable.

optimize_gbm_parameters(
  h2o_data,
  set,
  algorithm,
  metric,
  predictors,
  targets,
  runid
)

`h2o_data`	a list of H2OFrames
`set`	config object
`algorithm`	either `'gbm'` or `'xgboost'`
`metric`	evaluation metric
`predictors`	vector of explanatory variables
`targets`	label variable
`runid`	run ID