NADIA (version 0.4.2)

autotune_softImpute: Perform imputation using softImpute package

Description

Function use softImpute to impute missing data it works only with numeric data. Columns with categorical values are imputed by a selected function.

Usage

autotune_softImpute(
  df,
  percent_of_missing = NULL,
  col_type = NULL,
  col_0_1 = FALSE,
  cat_Fun = VIM::maxCat,
  lambda = 0,
  rank.max = 2,
  type = "als",
  thresh = 1e-05,
  maxit = 100,
  out_file = NULL
)

Value

Return one data.frame with imputed values.

Arguments

df

data.frame. Df to impute with column names and without target column.

percent_of_missing

numeric vector. Vector contatining percent of missing data in columns for example c(0,1,0,0,11.3,..)

col_type

Character vector with types of columns.

col_0_1

Decaid if add bonus column informing where imputation been done. 0 - value was in dataset, 1 - value was imputed. Default False. (Works only for returning one dataset).

cat_Fun

Function to impute categorical features. Default maxCat (mode). Can be every function with input one character vector and return atomic object.

lambda

nuclear-norm regularization parameter. If lambda=0, the algorithm reverts to "hardImpute", for which convergence is typically slower. If null lambda is set automatically at the highest possible values.

rank.max

This restricts the rank of the solution. Defoult 2 if set as NULL rank.max=min(dim(X))-1.

type

Chose of algoritm 'als' or 'svd . Defoult 'als'.

thresh

Threshold for convergence.

maxit

Maximum number of iterations.

out_file

Output log file location if file already exists log message will be added. If NULL no log will be produced.

Author

Trevor Hastie and Rahul Mazumder (2015).

Details

Function use algorithm base on matrix whats meaning if only one numeric column exists in dataset imputation algorithm don't work. In that case, this column will be imputed using a function for categorical columns. Because of this algorithm is working properly only with at least two numeric features in the dataset. To specify column type argument col_type is used so it's possible to forcefully use for example numeric factors in imputation. Action like this can led to errors and its not.

References

Trevor Hastie and Rahul Mazumder (2015). softImpute: Matrix Completion via Iterative Soft-Thresholded SVD. R package version 1.4. https://CRAN.R-project.org/package=softImpute

Examples

{
  raw_data <- data.frame(
    a = as.factor(sample(c("red", "yellow", "blue", NA), 1000, replace = TRUE)),
    b = as.integer(1:1000),
    c = as.factor(sample(c("YES", "NO", NA), 1000, replace = TRUE)),
    d = runif(1000, 1, 10),
    e = as.factor(sample(c("YES", "NO"), 1000, replace = TRUE)),
    f = as.factor(sample(c("male", "female", "trans", "other", NA), 1000, replace = TRUE)))

  # Prepering col_type
  col_type <- c("factor", "integer", "factor", "numeric", "factor", "factor")

  percent_of_missing <- 1:6
  for (i in percent_of_missing) {
    percent_of_missing[i] <- 100 * (sum(is.na(raw_data[, i])) / nrow(raw_data))
  }


  imp_data <- autotune_softImpute(raw_data, percent_of_missing, col_type)

  # Check if all missing value was imputed
  sum(is.na(imp_data)) == 0
  # TRUE
}