K Best Feature Selector#
A supervised feature selector that picks the top K ranked features returned by a learner implementing the RanksFeatures interface.
Note: The default feature ranking base learner is a fully-grown decision tree.
Interfaces: Transformer, Stateful, Persistable
Data Type Compatibility: Depends on the base learner
Parameters#
| # | Param | Default | Type | Description |
|---|---|---|---|---|
| 1 | k | int | The maximum number of features to select from the dataset. | |
| 2 | scorer | Auto | RanksFeatures | The base feature scorer. |
Additional Methods#
Return the final importances of the selected feature columns:
public importances() : ?array
Example#
use Rubix\ML\Transformers\KBestFeatureSelector;
use Rubix\ML\Regressors\GradientBoost;
$transformer = new KBestFeatureSelector(10, new GradientBoost());