assembleData: Assemble train and test data for model building with EWStools

In jknowles/EWStools: Tools for automating the testing and evaluation of education early warning system models

Description

Assemble train and test data for model building with EWStools

Usage

assembleData(data, class, p, predvars, classification = TRUE,
  keepNA = FALSE, ...)

Arguments

data	a dataframe that the user would like to split into training and sample sets
class	character value of the name of the dependent variable
p	the proportion of data to be placed into a training set
predvars	a character vector of the names of predictor variables
classification	Is the training set for a classification problem or not? Default is TRUE.
keepNA	Should missing values be preserved in the data sets? Logical. Default is FALSE.
...	additional arguments to be passed to assembleData

Details

To return a 3-way split with a validation set, use the pvalid argument.If classification is set to true the "class" component of the list will be forced to a factor for being fed into the train routine.

Value

A list of lists with the following items:

traindata

a list with a dataframe of the predictor matrix called preds, and the class called class

testdata

a list with a dataframe of the predictor matrix called preds, and the class called class

Note

Built on the createDataPartition function in the caret package.

jknowles/EWStools documentation built on May 19, 2019, 11:42 a.m.