Scikit-Learn Imputers
The following Scikit-Learn module types are missing value imputation methods.
Modules
KNNImputer
Scikit-Learn module documentation: KNNImputer
Arguments
| Description | Zero2Neuro Argument | Possible Values | Default Value | Scikit-Learn Parameter |
|---|---|---|---|---|
| Number of nearest neighbors for imputation | --skl_n_neighbors | Integer ≥ 1 | None | n_neighbors |
| Weight function for imputation | --skl_weights | uniform | distance | None | weights |
| Distance metric | --skl_metric | String; see sklearn docs | None | metric |
MissingIndicator
Scikit-Learn module documentation: MissingIndicator
Arguments
| Description | Zero2Neuro Argument | Possible Values | Default Value | Scikit-Learn Parameter |
|---|---|---|---|---|
| Which features to mark as missing | --skl_features | missing-only | all | None | features |
SimpleImputer
Scikit-Learn module documentation: SimpleImputer
Arguments
| Description | Zero2Neuro Argument | Possible Values | Default Value | Scikit-Learn Parameter |
|---|---|---|---|---|
| Strategy for imputing missing values | --skl_strategy | mean | median | most_frequent | constant | None | strategy |
Fill value when using constant strategy | --skl_fill_value | String | None | fill_value |