Travis build status Coverage status CRAN_Status_Badge Downloads

embed is a package that contains extra steps for the recipes package for embedding categorical predictors into one or more numeric columns. All of the preprocessing methods are supervised.

These steps are contained in a separate package because the package dependencies, rstanarm, lme4, and keras, are fairly heavy.

The steps included are:

  • step_lencode_glm, step_lencode_bayes, and step_lencode_mixed estimate the effect of each of the factor levels on the outcome and these estimates are used as the new encoding. The estimates are estimated by a generalized linear model. This step can be executed without pooling (via glm) or with partial pooling (stan_glm or lmer). Currently implemented for numeric and two-class outcomes.

  • step_embed uses keras::layer_embedding to translate the original C factor levels into a set of D new variables (< C). The model fitting routine optimizes which factor levels are mapped to each of the new variables as well as the corresponding regression coefficients (i.e., neural network weights) that will be used as the new encodings.

Some references for these methods are:

To install the package:

install.packages("embed")

## for development version:
require("devtools")
install_github("tidymodels/embed")