Applies target encoding to a given dataset

# S4 method for H2OTargetEncoderModel
h2o.transform(
  model,
  data,
  data_leakage_handling = NULL,
  use_blending = NULL,
  inflection_point = -1,
  smoothing = -1,
  noise = -1,
  seed = -1
)

Arguments

model

A trained model representing the transformation strategy

data

An H2OFrame with data to be transformed

data_leakage_handling

Handling of data leakage. Available options are : ["None", "LeaveOneOut", "KFold"]. Defaults to "None".

use_blending

Use blending during the transformation. Respects model settings when not set.

inflection_point

Blending parameter. Only effective when blending is enabled. By default, model settings are respected, if not overridden by this setting.

smoothing

Blending parameter. Only effective when blending is enabled. By default, model settings are respected, if not overridden by this setting.

noise

An amount of random noise added to the encoding. This helps prevent overfitting. Defaults to 0.01 * range of response.

seed

A random seed used to generate draws from the uniform distribution for random noise. Defaults to -1.

Value

Returns an H2OFrame object with data transformed.