RemixAutoML (version 0.5.0)

LimeModel: LimeModel to build a lime model

Description

LimeModel to build a lime model for prediction explanations in this package#'

Usage

LimeModel(
  data,
  Model = NULL,
  Bins = 10,
  ModelType = "xgboost",
  NThreads = parallel::detectCores(),
  MaxMem = "32G",
  ModelPath = NULL,
  ModelID = NULL
)

Arguments

data

Supply a training data set. This data set should be the data right before it gets converted to an h2o, catboost, or xgboost data object.

Model

Supply the model returned from training with the Auto__() functions.

Bins

Number of bins for discretizing numeric features

ModelType

Select from xgboost, h2o, and catboost

NThreads

Number of CPU threads

MaxMem

For use with H2O models. E.g. set to "28G"

ModelPath

Set to the path where your ML model is saved

ModelID

ID used to identify your ML model

Value

Model for utilizing lime

See Also

Other Model Evaluation and Interpretation: AutoLimeAid(), EvalPlot(), ML_EvalPlots(), ParDepCalPlots(), ROCPlot(), RedYellowGreen(), VI_Plot(), threshOptim()