hex.tree

Class DTree.Split

java.lang.Object

water.Iced

hex.tree.DTree.Split

All Implemented Interfaces:

java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, water.Freezable

Enclosing class:

DTree

public static class DTree.Split
extends water.Iced

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
int	_bin
int	_col

Constructor Summary

Constructors

Constructor and Description
Split(int col, int bin, DHistogram.NASplitDir nasplit, water.util.IcedBitSet bs, byte equal, double se, double se0, double se1, double n0, double n1, double p0, double p1, double tree_p0, double tree_p1)
Split(int col, int bin, DHistogram.NASplitDir nasplit, water.util.IcedBitSet bs, byte equal, double se, double se0, double se1, double n0, double n1, double p0, double p1, double tree_p0, double tree_p1, double p0Treat, double p0Contr, double p1Treat, double p1Contr, double n0Treat, double n0Contr, double n1Treat, double n1Contr, double upliftGainBefore, double upliftGain)

Method Summary

Modifier and Type	Method and Description
int	bin()
int	col()
double	n0()
double	n1()
DHistogram.NASplitDir	naSplitDir()
Constraints	nextLevelConstraints(Constraints currentConstraints, int way, double splat, SharedTreeModel.SharedTreeParameters parms)
DHistogram[]	nextLevelHistos(DHistogram[] currentHistos, int way, double splat, SharedTreeModel.SharedTreeParameters parms, Constraints cs, BranchInteractionConstraints bcs) Prepare children histograms, one per column.
double	pre_split_se()
double	preSplitUpliftGain()
double	se()
java.lang.String	toString()
double	upliftGain()

Methods inherited from class water.Iced

asBytes, clone, copyOver, frozenType, read, readExternal, readJSON, reloadFromBytes, toJsonBytes, toJsonString, write, writeExternal, writeJSON

Methods inherited from class java.lang.Object

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

Field Detail

_col

public final int _col

_bin

public final int _bin

Constructor Detail

Split

public Split(int col,
             int bin,
             DHistogram.NASplitDir nasplit,
             water.util.IcedBitSet bs,
             byte equal,
             double se,
             double se0,
             double se1,
             double n0,
             double n1,
             double p0,
             double p1,
             double tree_p0,
             double tree_p1)

Split

public Split(int col,
             int bin,
             DHistogram.NASplitDir nasplit,
             water.util.IcedBitSet bs,
             byte equal,
             double se,
             double se0,
             double se1,
             double n0,
             double n1,
             double p0,
             double p1,
             double tree_p0,
             double tree_p1,
             double p0Treat,
             double p0Contr,
             double p1Treat,
             double p1Contr,
             double n0Treat,
             double n0Contr,
             double n1Treat,
             double n1Contr,
             double upliftGainBefore,
             double upliftGain)

Method Detail

pre_split_se

public final double pre_split_se()

se

public final double se()

preSplitUpliftGain

public final double preSplitUpliftGain()

upliftGain

public final double upliftGain()

col

public final int col()

bin

public final int bin()

naSplitDir

public final DHistogram.NASplitDir naSplitDir()

n0

public final double n0()

n1

public final double n1()

nextLevelHistos

public DHistogram[] nextLevelHistos(DHistogram[] currentHistos,
                                    int way,
                                    double splat,
                                    SharedTreeModel.SharedTreeParameters parms,
                                    Constraints cs,
                                    BranchInteractionConstraints bcs)

Prepare children histograms, one per column. Typically, histograms are created with a level-dependent binning strategy. For the histogram of the current split decision, the children histograms are left/right range-adjusted. Any histgoram can null if there is no point in splitting further (such as there's fewer than min_row elements, or zero error in the response column). Return an array of DHistograms (one per column), which are bounded by the split bin-limits. If the column has constant data, or was not being tracked by a prior DHistogram (for being constant data from a prior split), then that column will be null in the returned array.

Parameters:

currentHistos - Histograms for all applicable columns computed for the previous split finding process

way - 0 (left) or 1 (right)

splat - Split point for previous split (if applicable)

parms - user-given parameters (will use nbins, min_rows, etc.)

Returns:

Array of histograms to be used for the next level of split finding

nextLevelConstraints

public Constraints nextLevelConstraints(Constraints currentConstraints,
                                        int way,
                                        double splat,
                                        SharedTreeModel.SharedTreeParameters parms)

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object