org.apache.spark.ml.tree

Class EnsembleModelReadWrite

Object

org.apache.spark.ml.tree.EnsembleModelReadWrite

public class EnsembleModelReadWrite
extends Object

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	EnsembleModelReadWrite.EnsembleNodeData Info for one Node in a tree ensemble
static class	EnsembleModelReadWrite.EnsembleNodeData$

Constructor Summary

Constructors

Constructor and Description
EnsembleModelReadWrite()

Method Summary

Modifier and Type	Method and Description
static scala.Tuple3<org.apache.spark.ml.util.DefaultParamsReader.Metadata,scala.Tuple2<org.apache.spark.ml.util.DefaultParamsReader.Metadata,Node>[],double[]>	loadImpl(String path, SparkSession sql, String className, String treeClassName) Helper method for loading a tree ensemble from disk.
static <M extends Params> void	saveImpl(M instance, String path, SparkSession sql, org.json4s.JsonAST.JObject extraMetadata) Helper method for saving a tree ensemble to disk.

Methods inherited from class Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

EnsembleModelReadWrite

public EnsembleModelReadWrite()

Method Detail

saveImpl

public static <M extends Params> void saveImpl(M instance,
                                               String path,
                                               SparkSession sql,
                                               org.json4s.JsonAST.JObject extraMetadata)

Helper method for saving a tree ensemble to disk.

Parameters:

instance - Tree ensemble model

path - Path to which to save the ensemble model.

extraMetadata - Metadata such as numFeatures, numClasses, numTrees.

sql - (undocumented)

loadImpl

public static scala.Tuple3<org.apache.spark.ml.util.DefaultParamsReader.Metadata,scala.Tuple2<org.apache.spark.ml.util.DefaultParamsReader.Metadata,Node>[],double[]> loadImpl(String path,
                                                                                                                                                                               SparkSession sql,
                                                                                                                                                                               String className,
                                                                                                                                                                               String treeClassName)

Helper method for loading a tree ensemble from disk. This reconstructs all trees, returning the root nodes.

Parameters:

path - Path given to saveImpl

className - Class name for ensemble model type

treeClassName - Class name for tree model type in the ensemble

sql - (undocumented)

Returns:

(ensemble metadata, array over trees of (tree metadata, root node)), where the root node is linked with all descendents

See Also:

saveImpl for how the model was saved