MakeTrainTest: Data-preprocessing and provide useful data for fitting

In SwanCygnus/predata: What the Package Does (One Line, Title Case)

Description

This function can arrange the data (Move the outcome to the left), standardize the training data and test data using training data's information. You can also choose whether you transform the outcome with Box-Cox transformation or not. If you transform the outcome, it also returns mean value outcome_mean, standard deviation outcome_sd and Box-Cox transformation parameter lambda

Usage

MakeTrainTest(
  outcome_name,
  features_index,
  data_train,
  data_test,
  y_method = FALSE,
  boxcox = FALSE
)

Arguments

outcome_name	The outcome's name (Strings)
features_index	The index of columns you use as features
data_train	Training data
data_test	Test data
y_method	Decide how outcome will be standardized. y_method = "scale": scale the outcome by dividing with it's standard deviation (y / sd(y)). y_method = "center": centering the outcome (y - mean(y)). c("scale", "center"): standardize the outcome ((y - mean(y)) / sd(y)). If FALSE, nothing is applied to the outcome.
boxcox	If TRUE, return transformed outcome with Box-Cox transformation using car package.

Value

A list that has scaled data and separate data in features and outcome as matrix.

train	Standardized training data
test	Standardized test data
train_x	Drop outcome from train
train_y	outcome of training data (if y_method was specified, It'll be a transformed value).
test_x	Drop outcome from test
test_y	Outcome of test data (not changed)
outcome_mean	A value used for centering outcome (mean of outcome in training data)
outcome_sd	A value used for scaling outcome (standard deviation of outcome in test data)

SwanCygnus/predata documentation built on July 1, 2020, 12:09 a.m.