train: Train a model encapsulated within a tundraContainer.
In syberia/tundra: Standardized container format for deployment of classifiers.

Description Usage Arguments Value

Train a model encapsulated within a tundraContainer.

1	train(dataframe, train_args = list(), verbose = FALSE, munge = TRUE)

`dataframe`	data.frame. The dataset to train the model on. This will be preprocessed with the `tundraContainer`'s `munge_procedure` and then passed as the first argument to the `tundraContainer`'s `train_function`.
`train_args`	list. A list of arguments to make available to the `tundraContainer`'s `train_function` through use of the `input` keyword. See the examples.
`verbose`	logical. Either `TRUE` or `FALSE`, by default the latter. If `TRUE`, then output produced by running the `munge_procedure` or the `train_function` will not be silenced.
`munge`	logical. Either `TRUE` or `FALSE`, by default the former. If `FALSE`, the `munge_procedure` provided to the container during initialization will be assumed to have been trained, and the `dataframe` provided will not be run through it.

The value returned by the tundraContainer's train_function. Since the train_function has side effects on the container as its primary purpose, this can usually be invisible(NULL).

syberia/tundra documentation built on May 30, 2019, 10:44 p.m.