train_test {lessR}

Create Training and Testing Data

Description

Given a data frame, create a list of either two components, train and test, or four components, for training and testing data: train_x, train_y, test_x, and test_y.

Usage

train_test(data, response=NULL, p_train=0.75, seed=NULL, matrix_out=FALSE)

Arguments

data

Data frame that contains the variables.

response

Optional name of the response variable of the response values.

p_train

Percentage of the input data frame to be retained for training.

seed

Set to a usually odd value to reproduce results.

matrix_out

If TRUE then output data structures as matrices instead of data frames.

Details

From the input data frame create training and testing data frames. If the response is specified, create four component data frames with x and y variables separated. Otherwise create two component data frames, train and test.

Author(s)

David W. Gerbing (Portland State University; [e-mail])

Examples

d <- Read("Employee")

# create four component data frames that separate the response variable, y,
#   from predictor variables, X: train_x, train_y, test_x, and test_y
out <- train_test(d, response=Salary)
names(out)
# then can copy to regular data frames apart from the list output structure
X_train <- out$train_x
y_train <- out$train_y
X_test <- out$test_x
y_test <- out$test_y

# create two component data frames, train and test, which retain all
#   variables for the model in the same data frame
out <- train_test(d)
names(out)
# then can copy to regular data frames apart from the list output structure
d_train <- out$train
d_test <- out$test