encode_lowrank {categoryEncodings}

Encode a given factor variable using low rank encoding

Description

Transforms the original design matrix using a low rank encoding.

Usage

encode_lowrank(X, fact, keep_factor = FALSE, encoding_only = FALSE)

Arguments

X

The data.frame/data.table to transform.

fact

The factor variable to encode by - either a positive integer specifying the column number, or the name of the column.

keep_factor

Whether to keep the original factor column(defaults to **FALSE**).

encoding_only

Whether to return the full transformed dataset or only the new columns. Defaults to FALSE and returns the full dataset.

Details

Uses the method from Johannemann et al.(2019) 'Sufficient Representations for Categorical Variables' - Low rank.

Value

A new data.table X which contains the new columns and optionally the old factor.

Examples

design_mat <- cbind( data.frame( matrix(rnorm(5*100),ncol = 5) ),
                     sample( sample(letters, 10), 100, replace = TRUE)
                     )
colnames(design_mat)[6] <- "factor_var"

encode_lowrank(X = design_mat, fact = "factor_var", keep_factor = FALSE)