In pmartR/peppuR:

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

After creating an as.MLinput() object, the next phase in the peppuR pipeline involves common preprocessing steps such as:

1. Handling missing values
2. Correlation filtering
3. Near-zero variance filtering
4. Univariate feature selection

Since we have no missing data, we'll proceed into correlation filtering which utilizes Max Kuhn's caret package. In general we use a correlation matrix based approach with the peppuR function univariate_feature_selection()

library(peppuR)
data("single_source")
single_source_peppuRobj <- univariate_feature_selection(single_source)

pmartR/peppuR documentation built on Jan. 17, 2020, 12:54 p.m.