org.apache.spark.mllib.tree.impurity

Class ImpurityAggregator

Object

org.apache.spark.mllib.tree.impurity.ImpurityAggregator

All Implemented Interfaces:

java.io.Serializable

Direct Known Subclasses:

EntropyAggregator, GiniAggregator, VarianceAggregator

public abstract class ImpurityAggregator
extends Object
implements scala.Serializable

Interface for updating views of a vector of sufficient statistics, in order to compute impurity from a sample. Note: Instances of this class do not hold the data; they operate on views of the data.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
ImpurityAggregator(int statsSize)

Method Summary

Methods

Modifier and Type	Method and Description
abstract ImpurityCalculator	getCalculator(double[] allStats, int offset) Get an ImpurityCalculator for a (node, feature, bin).
void	merge(double[] allStats, int offset, int otherOffset) Merge the stats from one bin into another.
int	statsSize()
abstract void	update(double[] allStats, int offset, double label, double instanceWeight) Update stats for one (node, feature, bin) with the given label.

Methods inherited from class Object

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

Constructor Detail

ImpurityAggregator

public ImpurityAggregator(int statsSize)

Method Detail

statsSize

public int statsSize()

merge

public void merge(double[] allStats,
         int offset,
         int otherOffset)

Merge the stats from one bin into another.

Parameters:

allStats - Flat stats array, with stats for this (node, feature, bin) contiguous.

offset - Start index of stats for (node, feature, bin) which is modified by the merge.

otherOffset - Start index of stats for (node, feature, other bin) which is not modified.

update

public abstract void update(double[] allStats,
          int offset,
          double label,
          double instanceWeight)

Update stats for one (node, feature, bin) with the given label.

Parameters:

allStats - Flat stats array, with stats for this (node, feature, bin) contiguous.

offset - Start index of stats for this (node, feature, bin).

getCalculator

public abstract ImpurityCalculator getCalculator(double[] allStats,
                               int offset)

Get an ImpurityCalculator for a (node, feature, bin).

Parameters:

allStats - Flat stats array, with stats for this (node, feature, bin) contiguous.

offset - Start index of stats for this (node, feature, bin).