org.apache.ignite.ml.tree.data

Class TreeDataIndex

java.lang.Object

org.apache.ignite.ml.tree.data.TreeDataIndex

public class TreeDataIndex
extends Object

Index for representing sorted dataset rows for each features. It may be reused while decision tree learning at several levels through filter method.

Constructor Summary

Constructors

Constructor and Description
TreeDataIndex(double[][] features, double[] labels) Constructs an instance of TreeDataIndex.

Method Summary

Modifier and Type	Method and Description
int	columnsCount()
double	featureInSortedOrder(int k, int featureId) Returns feature value for kth order statistic for target feature.
double[]	featuresInSortedOrder(int k, int featureId) Returns vector of original features for kth order statistic for target feature.
TreeDataIndex	filter(TreeFilter filter) Creates projection of current index in according to TreeFilter.
double	labelInSortedOrder(int k, int featureId) Returns label for kth order statistic for target feature.
int	rowsCount()

Methods inherited from class java.lang.Object

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

Constructor Detail

TreeDataIndex

public TreeDataIndex(double[][] features,
                     double[] labels)

Constructs an instance of TreeDataIndex.

Parameters:

features - Features.

labels - Labels.

Method Detail

labelInSortedOrder

public double labelInSortedOrder(int k,
                                 int featureId)

Returns label for kth order statistic for target feature.

Parameters:

k - K.

featureId - Feature id.

Returns:

Label value.

featuresInSortedOrder

public double[] featuresInSortedOrder(int k,
                                      int featureId)

Returns vector of original features for kth order statistic for target feature.

Parameters:

k - K.

featureId - Feature id.

Returns:

Features vector.

featureInSortedOrder

public double featureInSortedOrder(int k,
                                   int featureId)

Returns feature value for kth order statistic for target feature.

Parameters:

k - K.

featureId - Feature id.

Returns:

Feature value.

filter

public TreeDataIndex filter(TreeFilter filter)

Creates projection of current index in according to TreeFilter.

Parameters:

filter - Filter.

Returns:

Projection of current index onto smaller index in according to rows filter.

rowsCount

public int rowsCount()

Returns:

count of rows in current index.

columnsCount

public int columnsCount()

Returns:

count of columns in current index.