[source] # K Best Feature Selector A supervised feature selector that picks the top K ranked features returned by a learner implementing the [RanksFeatures](../ranks-features.md) interface. !!! note The default feature ranking base learner is a fully-grown decision tree. **Interfaces:** [Transformer](api.md#transformer), [Stateful](api.md#stateful), [Persistable](../persistable.md) **Data Type Compatibility:** Depends on the base learner ## Parameters | # | Name | 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: ``` php public importances() : ?array ``` ## Example ```php use Rubix\ML\Transformers\KBestFeatureSelector; use Rubix\ML\Regressors\GradientBoost; $transformer = new KBestFeatureSelector(10, new GradientBoost()); ```