PreprocessorConfig

Properties

complete_cases

booleandefault false

Retain only complete cases.

remove_features_thres

number | nulldefault null≤ 1> 0

Remove features missing in >= this fraction of cases.

remove_cases_thres

number | nulldefault null≤ 1> 0

Remove cases missing >= this fraction of features.

missingness

booleandefault false

Add a boolean missingness indicator per feature with NAs.

impute

booleandefault false

Impute missing values.

impute_type

stringdefault "missRanger"

Imputation method.

one of"missRanger""micePMM""meanMode"

impute_discrete

stringdefault "get_mode"

Function name to impute discrete features.

impute_continuous

stringdefault "mean"

Function name to impute continuous features.

integer2factor

booleandefault false

Convert integers to factors.

integer2numeric

booleandefault false

Convert integers to numeric.

logical2factor

booleandefault false

Convert logicals to factors.

logical2numeric

booleandefault false

Convert logicals to numeric.

numeric2factor

booleandefault false

Convert numeric to factors.

numeric2factor_levels

array | nulldefault null≥ 1 items

Factor levels for numeric2factor.

numeric_cut_n

integerdefault 0≥ 0

Cut numeric features into this many bins (0 = off).

numeric_cut_labels

booleandefault false

Use labels for numeric_cut bins.

numeric_quant_n

integerdefault 0≥ 0

Cut numeric features into this many quantile bins (0 = off).

numeric_quant_NAonly

booleandefault false

Quantile-cut only features with NAs.

unique_len2factor

integerdefault 0≥ 0

Convert features with <= this many unique values to factors (0 = off).

character2factor

booleandefault false

Convert character features to factors.

factorNA2missing

booleandefault false

Convert factor NAs to a 'missing' level.

factorNA2missing_level

stringdefault "missing"

Level name for factorNA2missing.

factor2integer

booleandefault false

Convert factors to integers.

factor2integer_startat0

booleandefault true

factor2integer starts at 0.

scale

booleandefault false

Scale features.

center

booleandefault false

Center features.

remove_constants

booleandefault false

Remove constant features.

remove_constants_skip_missing

booleandefault true

Ignore missing values when detecting constants.

remove_duplicates

booleandefault false

Remove duplicate cases.

remove_features

array | nulldefault null≥ 1 items

Names of features to remove.

one_hot

booleandefault false

One-hot encode factors.

add_date_features

booleandefault false

Add date-derived features.

date_features

string[]default ["weekday","month","year"]≥ 1 items

Date features to add.

add_holidays

booleandefault false

Add a holiday indicator feature.

exclude

array | nulldefault null≥ 1 items

Column indices to exclude from preprocessing.

impute_missRanger_params

object

Parameters passed to missRanger (e.g. pmm.k, maxiter, num.trees).

scale_centers

object | null

scale_coefficients

object | null

one_hot_levels

object | null

Relationships