hex.singlenoderf

Class VariableImportance.TreeMeasures<T extends VariableImportance.TreeMeasures>

java.lang.Object

water.Iced

hex.singlenoderf.VariableImportance.TreeMeasures<T>

All Implemented Interfaces:

java.lang.Cloneable, Freezable

Direct Known Subclasses:

VariableImportance.TreeVotes

Enclosing class:

VariableImportance

public abstract static class VariableImportance.TreeMeasures<T extends VariableImportance.TreeMeasures>
extends Iced

A simple holder for set of different tree measurements.

Field Summary

Fields

Modifier and Type	Field and Description
protected long[]	_nrows Number of processed row per tree.
protected int	_ntrees Actual number of trees which votes are stored in this object

Constructor Summary

Constructors

Constructor and Description
VariableImportance.TreeMeasures(int initialCapacity)
VariableImportance.TreeMeasures(long[] nrows, int ntrees)

Method Summary

Methods

Modifier and Type	Method and Description
double[]	accuracy()
abstract double	accuracy(int tidx) Returns a list of accuracies per tree.
abstract T	append(T t)
abstract double[]	imp(T right) Compute variable importance with respect to given votes.
int	npredictors() Returns number of voting predictors
long[]	nrows() Returns number of rows which were used during voting per individual tree.

Methods inherited from class water.Iced

clone, frozenType, init, newInstance, read, toDocField, write, writeJSON, writeJSONFields

Methods inherited from class java.lang.Object

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

Field Detail

_ntrees

protected int _ntrees

Actual number of trees which votes are stored in this object

_nrows

protected long[] _nrows

Number of processed row per tree.

Constructor Detail

VariableImportance.TreeMeasures

public VariableImportance.TreeMeasures(int initialCapacity)

VariableImportance.TreeMeasures

public VariableImportance.TreeMeasures(long[] nrows,
                               int ntrees)

Method Detail

nrows

public final long[] nrows()

Returns number of rows which were used during voting per individual tree.

npredictors

public final int npredictors()

Returns number of voting predictors

accuracy

public abstract double accuracy(int tidx)

Returns a list of accuracies per tree.

accuracy

public final double[] accuracy()

imp

public abstract double[] imp(T right)

Compute variable importance with respect to given votes. The given T object represents correct votes. This object represents votes over shuffled data.

Parameters:

right - individual tree measurements performed over not shuffled data.

Returns:

computed importance and standard deviation

append

public abstract T append(T t)