org.apache.ignite.ml.math.impls.vector

Class AbstractReadOnlyVector

java.lang.Object

org.apache.ignite.ml.math.impls.vector.AbstractVector

org.apache.ignite.ml.math.impls.vector.AbstractReadOnlyVector

All Implemented Interfaces:

Externalizable, Serializable, Destroyable, MetaAttributes, StorageOpsMetrics, Vector

Direct Known Subclasses:

ConstantVector, RandomVector

public abstract class AbstractReadOnlyVector
extends AbstractVector

This class provides a helper implementation of the read-only implementation of Vector interface to minimize the effort required to implement it. Subclasses may override some of the implemented methods if a more specific or optimized implementation is desirable.

See Also:

Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from interface org.apache.ignite.ml.math.Vector

Vector.Element

Constructor Summary

Constructors

Constructor and Description
AbstractReadOnlyVector()
AbstractReadOnlyVector(VectorStorage sto)

Method Summary

Modifier and Type	Method and Description
void	compute(int idx, IgniteIntDoubleToDoubleBiFunction f) Replace vector entry with value oldVal at i with result of computing f(i, oldVal).
Vector	copy() Creates new copy of this vector.
Matrix	cross(Vector vec) Gets the cross product of this vector and the other vector.
Vector	divide(double x) Creates new vector containing values from this vector divided by the argument.
Vector	logNormalize() Creates new vector containing the log(1 + entry) / L_2 norm values of this vector.
Vector	logNormalize(double power) Creates new vector with a normalized value calculated as log_power(1 + entry) / L_power norm.
Vector	map(IgniteBiFunction<Double,Double,Double> fun, double y) Maps all elements of this vector by applying given function to each element with a constant second parameter y.
Vector	map(IgniteDoubleFunction<Double> fun) Maps all values in this vector through a given function.
Vector	map(Vector vec, IgniteBiFunction<Double,Double,Double> fun) Maps all values in this vector through a given function.
Vector	times(double x) Gets a new vector that contains product of each element and the argument.
Matrix	toMatrix(boolean rowLike) Converts this vector into [N x 1] or [1 x N] matrix where N is this vector cardinality.
Matrix	toMatrixPlusOne(boolean rowLike, double zeroVal) Converts this vector into [N+1 x 1] or [1 x N+1] matrix where N is this vector cardinality

Methods inherited from class org.apache.ignite.ml.math.impls.vector.AbstractVector

all, allSpliterator, assign, assign, assign, assign, checkCardinality, checkCardinality, checkCardinality, checkIndex, destroy, dot, dotSelf, equals, foldMap, foldMap, get, getDistanceSquared, getElement, getLengthSquared, getMetaStorage, getStorage, getX, guid, hashCode, increment, incrementX, isArrayBased, isDense, isDistributed, isRandomAccess, isSequentialAccess, isZero, kNorm, makeElement, maxElement, maxValue, minElement, minus, minValue, nonZeroElements, nonZeroes, nonZeroSpliterator, normalize, normalize, plus, plus, readExternal, set, setStorage, setX, size, sort, storageGet, storageSet, sum, times, viewPart, writeExternal

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.ignite.ml.math.Vector

like, likeMatrix

Methods inherited from interface org.apache.ignite.ml.math.MetaAttributes

getAttribute, hasAttribute, removeAttribute, setAttribute

Constructor Detail

AbstractReadOnlyVector

public AbstractReadOnlyVector()

AbstractReadOnlyVector

public AbstractReadOnlyVector(VectorStorage sto)

Parameters:

sto - Storage.

Method Detail

cross

public Matrix cross(Vector vec)

Gets the cross product of this vector and the other vector.

Specified by:

cross in interface Vector

Overrides:

cross in class AbstractVector

Parameters:

vec - Second vector.

Returns:

New matrix as a cross product of two vectors.

toMatrix

public Matrix toMatrix(boolean rowLike)

Converts this vector into [N x 1] or [1 x N] matrix where N is this vector cardinality.

Specified by:

toMatrix in interface Vector

Overrides:

toMatrix in class AbstractVector

Parameters:

rowLike - true for rowLike [N x 1], or false for column [1 x N] matrix.

Returns:

Newly created matrix.

toMatrixPlusOne

public Matrix toMatrixPlusOne(boolean rowLike,
                              double zeroVal)

Converts this vector into [N+1 x 1] or [1 x N+1] matrix where N is this vector cardinality. (0,0) element of this matrix will be zeroVal parameter.

Specified by:

toMatrixPlusOne in interface Vector

Overrides:

toMatrixPlusOne in class AbstractVector

Parameters:

rowLike - true for rowLike [N+1 x 1], or false for column [1 x N+1] matrix.

Returns:

Newly created matrix.

copy

public Vector copy()

Creates new copy of this vector.

Specified by:

copy in interface Vector

Overrides:

copy in class AbstractVector

Returns:

New copy vector.

logNormalize

public Vector logNormalize()

Creates new vector containing the log(1 + entry) / L_2 norm values of this vector.

Specified by:

logNormalize in interface Vector

Overrides:

logNormalize in class AbstractVector

Returns:

New vector.

logNormalize

public Vector logNormalize(double power)

Creates new vector with a normalized value calculated as log_power(1 + entry) / L_power norm.

Specified by:

logNormalize in interface Vector

Overrides:

logNormalize in class AbstractVector

Parameters:

power - The power to use. Must be > 1. Cannot be Double.POSITIVE_INFINITY.

Returns:

New vector

map

public Vector map(IgniteDoubleFunction<Double> fun)

Maps all values in this vector through a given function.

Specified by:

map in interface Vector

Overrides:

map in class AbstractVector

Parameters:

fun - Mapping function.

Returns:

This vector.

map

public Vector map(Vector vec,
                  IgniteBiFunction<Double,Double,Double> fun)

Maps all values in this vector through a given function. For this vector A, argument vector B and the function F this method maps every element x as: A(x) = F(A(x), B(x))

Specified by:

map in interface Vector

Overrides:

map in class AbstractVector

Parameters:

vec - Argument vector.

fun - Mapping function.

Returns:

This function.

map

public Vector map(IgniteBiFunction<Double,Double,Double> fun,
                  double y)

Maps all elements of this vector by applying given function to each element with a constant second parameter y.

Specified by:

map in interface Vector

Overrides:

map in class AbstractVector

Parameters:

fun - Mapping function.

y - Second parameter for mapping function.

Returns:

This vector.

divide

public Vector divide(double x)

Creates new vector containing values from this vector divided by the argument.

Specified by:

divide in interface Vector

Overrides:

divide in class AbstractVector

Parameters:

x - Division argument.

Returns:

New vector.

times

public Vector times(double x)

Gets a new vector that contains product of each element and the argument.

Specified by:

times in interface Vector

Overrides:

times in class AbstractVector

Parameters:

x - Multiply argument.

Returns:

New vector.

compute

public void compute(int idx,
                    IgniteIntDoubleToDoubleBiFunction f)

Replace vector entry with value oldVal at i with result of computing f(i, oldVal).

Specified by:

compute in interface Vector

Overrides:

compute in class AbstractVector

Parameters:

idx - Position.

f - Function used for replacing.