sparklyr (version 0.6.4)

ml_options: Options for Spark ML Routines

Description

Provide this object to the various Spark ML methods, to control certain facets of the model outputs produced.

Usage

ml_options(id.column = random_string("id"),
  response.column = random_string("response"),
  features.column = random_string("features"),
  output.column = random_string("output"), model.transform = NULL,
  only.model = FALSE, na.action = getOption("na.action", "na.omit"), ...)

Arguments

id.column

The name to assign to the generated id column.

response.column

The name to assign to the generated response column.

features.column

The name to assign to the generated features column.

output.column

The name to assign to the generated output column.

model.transform

An optional R function that accepts a Spark model and returns a Spark model. This can be used to supply optional Spark model fitting parameters not made available in the sparklyr APIs.

only.model

Boolean; should the Spark model object itself be returned without fitting the actual model? Useful for ml_one_vs_rest.

na.action

An R function, or the name of an R function, indicating how missing values should be handled.

...

Optional arguments, reserved for future expansion.