insights

Type Members

1. sealed abstract class NormType extends EnumEntry with Serializable

   

   Represents a kind of scaling to do on feature before computing importance

2. case class Normalizer(scaleFactors1: Array[Double], scaleFactors2: Array[Double], offset: Double, name: String) extends Serializable with Product

   

3. class RecordInsightsCorr extends BinaryEstimator[OPVector, OPVector, TextMap]

   

   Creates record level insights for model predictions.

   Creates record level insights for model predictions. Takes two inputs the first is the predictions to explain and the second feature vector fed into the model. Note that regression outputs must be converted into a vector with one column in order to use this estimator

4. class RecordInsightsLOCO[T <: Model[T]] extends UnaryTransformer[OPVector, TextMap] with RecordInsightsLOCOParams

   

   Creates record level insights for model predictions.

   Creates record level insights for model predictions. Takes the model to explain as a constructor argument. The input feature is the feature vector fed into the model.

   The map's contents are different regarding the value of the topKStrategy param (only for Binary Classification and Regression) : - If PositiveNegative, returns at most 2 * topK elements : the topK most positive and the topK most negative derived features based on the LOCO insight.For MultiClassification, the value is from the predicted class (i.e. the class having the highest probability) - If Abs, returns at most topK elements : the topK derived features having highest absolute value of LOCO score.

   Annotations

   @Experimental()

5. trait RecordInsightsLOCOParams extends Params

   

6. sealed abstract class TopKStrategy extends EnumEntry with Serializable

   

7. sealed abstract class VectorAggregationStrategy extends EnumEntry with Serializable