pre_impute_df: Impute a data frame

In emil: Evaluation of Modeling without Information Leakage

Description

This function imputes each column of data frames univariately with different functions depending on their class.

Usage

pre_impute_df(data, class_fun = list(numeric = function(x) median(x, na.rm =
  TRUE), integer = function(x) median(x, na.rm = TRUE)),
  default_fun = function(x) mode(x, na.rm = TRUE, allow_multiple = FALSE))

Arguments

data	Pre-processed data set with features in a data frame.
class_fun	List of functions to use for imputating specific feature classes.
default_fun	Function to use for imputation features of classes not listed in class_fun.
na.rm	Whether to remove missing values.

Author(s)

Christofer Bäcklin

Examples

x <- iris
x[sample(150, 3), 1] <- NA                                                                          
x[sample(150, 1), 3] <- NA
x[sample(150, 5), 5] <- NA
y <- gl(2, 75)
fold <- resample("holdout", y, nfold=1)[[1]]
data <- pre_split(x, y, fold) %>%
    pre_impute_df

emil documentation built on Aug. 1, 2018, 1:03 a.m.