org.apache.ignite.ml.tree.randomforest.data.impurity

Class ImpurityHistogramsComputer<S extends ImpurityComputer<BootstrappedVector,S>>

java.lang.Object

org.apache.ignite.ml.tree.randomforest.data.impurity.ImpurityHistogramsComputer<S>

Type Parameters:

S - Type of basic impurity computer for feature.

All Implemented Interfaces:

Serializable

Direct Known Subclasses:

GiniHistogramsComputer, MSEHistogramComputer

public abstract class ImpurityHistogramsComputer<S extends ImpurityComputer<BootstrappedVector,S>>
extends Object
implements Serializable

Class containing logic of aggregation impurity statistics within learning dataset.

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	ImpurityHistogramsComputer.NodeImpurityHistograms<S extends ImpurityComputer<BootstrappedVector,S>> Class represents per feature statistics for impurity computing.

Constructor Summary

Constructors

Constructor and Description
ImpurityHistogramsComputer()

Method Summary

Modifier and Type	Method and Description
Map<NodeId,ImpurityHistogramsComputer.NodeImpurityHistograms<S>>	aggregateImpurityStatistics(ArrayList<TreeRoot> roots, Map<Integer,BucketMeta> histMeta, Map<NodeId,TreeNode> nodesToLearn, Dataset<EmptyContext,BootstrappedDatasetPartition> dataset) Computes histograms for each feature.
protected abstract S	createImpurityComputerForFeature(int sampleId, BucketMeta meta) Creates impurity computer in according to specific algorithm based on random forest (for example GiniHistogram for classification).

Methods inherited from class java.lang.Object

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

Constructor Detail

ImpurityHistogramsComputer

public ImpurityHistogramsComputer()

Method Detail

aggregateImpurityStatistics

public Map<NodeId,ImpurityHistogramsComputer.NodeImpurityHistograms<S>> aggregateImpurityStatistics(ArrayList<TreeRoot> roots,
                                                                                                    Map<Integer,BucketMeta> histMeta,
                                                                                                    Map<NodeId,TreeNode> nodesToLearn,
                                                                                                    Dataset<EmptyContext,BootstrappedDatasetPartition> dataset)

Computes histograms for each feature.

Parameters:

roots - Random forest roots.

histMeta - Histograms meta.

nodesToLearn - Nodes to learn.

dataset - Dataset.

createImpurityComputerForFeature

protected abstract S createImpurityComputerForFeature(int sampleId,
                                                      BucketMeta meta)

Creates impurity computer in according to specific algorithm based on random forest (for example GiniHistogram for classification).

Parameters:

sampleId - Sample id.

meta - Bucket Meta.

Returns:

Impurity computer