impute_missing_data: Impute Missing Data

In KevinSee/QRFcapacity: Carrying Capacity with QRF

Impute Missing Data

Description

Filter out rows with too much missing data, impute the remaining missing values

Usage

impute_missing_data(
  data = NULL,
  covars = NULL,
  max_missing = 3,
  impute_vars = NULL,
  method = c("Hmisc", "missForest", "randomForestSRC"),
  ntree = 100,
  nk = 4,
  ...
)

Arguments

data	data.frame containing covars and any impute_vars as columns
covars	vector of covariate names to be imputed, which match column names in data
max_missing	integer value of the maximum number of columns per row that may have missing values. Rows with more than this number of missing covariates will be deleted.
impute_vars	additional covariates included in data that may be used for imputation
method	which method to use, either by creating random forests using the missForest package, or by predictive mean matching using the aregImpute function in the Hmisc package
ntree	how many trees to build if method = 'missForest' or method == 'randomForestSRC'
nk	how many knots to use in smoothing splines if using method = 'Hmisc'
...	other arguments to be passed to either missForest, aregImpute or impute functions

Value

data.frame containing non-missing and imputed data

Author(s)

Kevin See

KevinSee/QRFcapacity documentation built on Feb. 27, 2023, 3:57 p.m.