metadata

Type Members

1. class ClassificationModelMetadata extends RowModelMetadata

   

2. class ClusteringModelMetadata extends RowModelMetadata

   

3. case class DetailedTransformationSchema(inputFeatures: Seq[ColumnDef], outputFeatures: Seq[ColumnDef], identifier: String = "") extends TransformationSchema with Product with Serializable

   

   If the user provides us with the input feature descriptions, then in the Predictor we will be able to verify that the data set for prediction contains all the necessary columns at design-time (not currently implemented but should be easy to do).

4. class RegressionModelMetadata extends RowModelMetadata

   

5. trait RowModelMetadata extends IOMetadata

   

6. case class RowModelMetadataDefault(transformationSchema: TransformationSchema, sqlScorable: Boolean, sqlOutputFeatures: Seq[ColumnDef]) extends RowModelMetadata with Product with Serializable

   

7. class TransformationSchema extends AnyRef

   

   We need this information to figure out the output structure of the Predictor, e.g.

   We need this information to figure out the output structure of the Predictor, e.g. if the resultType is ClassificationResult then there are 3 output columns, if it is RealResult then there is one (double type) column. identifier is used as part of the output column names (we may change how this works in the future).