cc.mallet.types
Class SparseVector

java.lang.Object
  cc.mallet.types.SparseVector

All Implemented Interfaces:

ConstantMatrix, Vector, java.io.Serializable

Direct Known Subclasses:

FeatureVector, HashedSparseVector, IndexedSparseVector

public class SparseVector
extends java.lang.Object
implements ConstantMatrix, Vector, java.io.Serializable

A vector that allocates memory only for non-zero values. When you create a SparseVector, you pass in a list of indices. These are the only elements of the vector you will be allowed to change. The rest are fixed at 0. The interface to Sparse vector uses the concept of a location, which is an integer in the range 0..numLocations which can be mapped to the index (and value) of a non zero element of the vector. A SparseVector can be sparse or dense depending on whether or not an array if indices is specified at construction time. If the SparseVector is dense, the mapping from location to index is the identity mapping. The type of the value an element in a SparseVector (or FeatureVector) can be double or binary (0.0 or 1.0), depending on whether an array of doubles is specified at contruction time.

Author:

Andrew McCallum mccallum@cs.umass.edu

See Also:

Serialized Form

Field Summary

protected boolean	hasInfinite
protected int[]	indices If the vector is sparse, then both indices and values are sparse.
protected double[]	values

Constructor Summary

SparseVector()
An empty vector, with all zero values

SparseVector(Alphabet dict, PropertyList pl, boolean binary)

SparseVector(Alphabet dict, PropertyList pl, boolean binary, boolean growAlphabet)

SparseVector(double[] values)

SparseVector(double[] values, boolean copy)

SparseVector(int[] indices)

SparseVector(int[] indices, boolean copy)

SparseVector(int[] indices, boolean copy, boolean checkIndicesSorted)

SparseVector(int[] indices, boolean copy, boolean checkIndicesSorted, boolean removeDuplicates, boolean binary)

SparseVector(int[] indices, double[] values)

SparseVector(int[] indices, double[] values, boolean copy)

SparseVector(int[] indices, double[] values, boolean copy, boolean checkIndicesSorted)

SparseVector(int[] indices, double[] values, boolean copy, boolean checkIndicesSorted, boolean removeDuplicates)

SparseVector(int[] indices, double[] values, int capacity, int size, boolean copy, boolean checkIndicesSorted, boolean removeDuplicates)
If "indices" is null, the vector will be dense.

SparseVector(int[] indices, int capacity, int size, boolean copy, boolean checkIndicesSorted, boolean removeDuplicates)

SparseVector(int size, double fillValue)

Method Summary

double	absNorm()
void	addTo(double[] accumulator)
void	addTo(double[] accumulator, double scale)
void	arrayCopyFrom(double[] a) Copy values from an array into this vector.
int	arrayCopyFrom(double[] a, int startingArrayLocation) Copy values from an array starting at a particular location into this vector.
int	arrayCopyInto(double[] array, int startingArrayLocation) Copy the contents of this vector into an array starting at a particular location.
ConstantMatrix	cloneMatrix() CLONING
ConstantMatrix	cloneMatrixZeroed()
double	dotProduct(ConstantMatrix m)
double	dotProduct(DenseVector v)
double	dotProduct(double[] v) VECTOR OPERATIONS
double	dotProduct(SparseVector v)
double	extendedDotProduct(DenseVector v)
double	extendedDotProduct(SparseVector v)
int	getDimensions(int[] sizes)
int[]	getIndices()
int	getNumDimensions() ACCESSORS
double[]	getValues()
void	incrementValue(int index, double value) Increments this[index] by value.
int	indexAtLocation(int location)
double	infinityNorm()
boolean	isBinary()
boolean	isInfinite()
boolean	isNaN()
boolean	isNaNOrInfinite()
int	location(int index)
void	makeBinary()
void	makeNonBinary()
void	map(java.lang.reflect.Method f) Applies the method argument to each value in a non-binary vector.
int	numLocations()
double	oneNorm()
void	plusEqualsSparse(SparseVector v) For each index i that is present in this vector, set this[i] += v[i].
void	plusEqualsSparse(SparseVector v, double factor) For each index i that is present in this vector, set this[i] += factor * v[i].
void	print()
protected void	removeDuplicates(int numDuplicates)
void	setAll(double v) Sets every present index in the vector to v.
void	setValue(int index, double value) Sets the value at the given index.
void	setValueAtLocation(int location, double value) Sets the value at the given location.
int	singleIndex(int[] indices)
int	singleSize()
void	singleToIndices(int i, int[] indices)
double	singleValue(int i)
protected void	sortIndices()
void	timesEquals(double factor) Scale all elements by the same factor.
void	timesEqualsSparse(SparseVector v) For each index i that is present in this vector, set this[i] *= v[i].
void	timesEqualsSparse(SparseVector v, double factor) For each index i that is present in this vector, set this[i] *= factor * v[i].
void	timesEqualsSparseZero(SparseVector v, double factor) For each index i that is present in this vector, set this[i] *= factor * v[i].
java.lang.String	toString()
java.lang.String	toString(boolean onOneLine)
double	twoNorm()
double	value(int index)
double	value(int[] indices)
double	valueAtLocation(int location)
SparseVector	vectorAdd(SparseVector v, double scale)

Methods inherited from class java.lang.Object

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

Field Detail

indices

protected int[] indices

If the vector is sparse, then both indices and values are sparse. Indices into these arrays are called ``locations'' in the below. The indices[] array maps locations to indices of the (virtual) dense array that's being represented. value[] maps locations to values.

values

protected double[] values

hasInfinite

protected boolean hasInfinite

Constructor Detail

SparseVector

public SparseVector(int[] indices,
                    double[] values,
                    int capacity,
                    int size,
                    boolean copy,
                    boolean checkIndicesSorted,
                    boolean removeDuplicates)

If "indices" is null, the vector will be dense. If "values" is null, the vector will be binary. The capacity and size arguments are used by AugmentableFeatureVector.

SparseVector

public SparseVector(double[] values,
                    boolean copy)

SparseVector

public SparseVector(double[] values)

SparseVector

public SparseVector(int size,
                    double fillValue)

SparseVector

public SparseVector(int[] indices,
                    double[] values,
                    boolean copy,
                    boolean checkIndicesSorted,
                    boolean removeDuplicates)

SparseVector

public SparseVector(int[] indices,
                    double[] values)

SparseVector

public SparseVector(int[] indices,
                    double[] values,
                    boolean copy)

SparseVector

public SparseVector(int[] indices,
                    double[] values,
                    boolean copy,
                    boolean checkIndicesSorted)

SparseVector

public SparseVector(int[] indices,
                    boolean copy,
                    boolean checkIndicesSorted,
                    boolean removeDuplicates,
                    boolean binary)

SparseVector

public SparseVector(int[] indices,
                    int capacity,
                    int size,
                    boolean copy,
                    boolean checkIndicesSorted,
                    boolean removeDuplicates)

SparseVector

public SparseVector(int[] indices,
                    boolean copy,
                    boolean checkIndicesSorted)

SparseVector

public SparseVector(int[] indices,
                    boolean copy)

SparseVector

public SparseVector(int[] indices)

SparseVector

public SparseVector()

An empty vector, with all zero values

SparseVector

public SparseVector(Alphabet dict,
                    PropertyList pl,
                    boolean binary,
                    boolean growAlphabet)

SparseVector

public SparseVector(Alphabet dict,
                    PropertyList pl,
                    boolean binary)

Method Detail

isBinary

public boolean isBinary()

makeBinary

public void makeBinary()

makeNonBinary

public void makeNonBinary()

getNumDimensions

public int getNumDimensions()

ACCESSORS

Specified by:

getNumDimensions in interface ConstantMatrix

getDimensions

public int getDimensions(int[] sizes)

Specified by:

getDimensions in interface ConstantMatrix

getIndices

public int[] getIndices()

getValues

public double[] getValues()

numLocations

public int numLocations()

Specified by:

numLocations in interface ConstantMatrix

location

public int location(int index)

Specified by:

location in interface ConstantMatrix

valueAtLocation

public double valueAtLocation(int location)

Specified by:

valueAtLocation in interface ConstantMatrix

indexAtLocation

public int indexAtLocation(int location)

Specified by:

indexAtLocation in interface ConstantMatrix

value

public double value(int[] indices)

Specified by:

value in interface ConstantMatrix

value

public double value(int index)

Specified by:

value in interface Vector

addTo

public void addTo(double[] accumulator,
                  double scale)

addTo

public void addTo(double[] accumulator)

singleIndex

public int singleIndex(int[] indices)

Specified by:

singleIndex in interface ConstantMatrix

singleToIndices

public void singleToIndices(int i,
                            int[] indices)

Specified by:

singleToIndices in interface ConstantMatrix

singleValue

public double singleValue(int i)

Specified by:

singleValue in interface ConstantMatrix

singleSize

public int singleSize()

Specified by:

singleSize in interface ConstantMatrix

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

toString

public java.lang.String toString(boolean onOneLine)

cloneMatrix

public ConstantMatrix cloneMatrix()

CLONING

Specified by:

cloneMatrix in interface ConstantMatrix

cloneMatrixZeroed

public ConstantMatrix cloneMatrixZeroed()

plusEqualsSparse

public void plusEqualsSparse(SparseVector v)

For each index i that is present in this vector, set this[i] += v[i]. If v has indices that are not present in this, these are just ignored.

plusEqualsSparse

public void plusEqualsSparse(SparseVector v,
                             double factor)

For each index i that is present in this vector, set this[i] += factor * v[i]. If v has indices that are not present in this, these are just ignored.

timesEqualsSparse

public void timesEqualsSparse(SparseVector v)

For each index i that is present in this vector, set this[i] *= v[i]. If v has indices that are not present in this, these are just ignored.

timesEqualsSparse

public void timesEqualsSparse(SparseVector v,
                              double factor)

For each index i that is present in this vector, set this[i] *= factor * v[i]. If v has indices that are not present in this, these are just ignored.

timesEqualsSparseZero

public void timesEqualsSparseZero(SparseVector v,
                                  double factor)

For each index i that is present in this vector, set this[i] *= factor * v[i]. If v has indices that are not present in this, these are set to zero

timesEquals

public void timesEquals(double factor)

Scale all elements by the same factor.

incrementValue

public void incrementValue(int index,
                           double value)
                    throws java.lang.IllegalArgumentException

Increments this[index] by value.

Throws:

java.lang.IllegalArgumentException - If index is not present.

setAll

public void setAll(double v)

Sets every present index in the vector to v.

setValue

public void setValue(int index,
                     double value)
              throws java.lang.IllegalArgumentException

Sets the value at the given index.

Throws:

java.lang.IllegalArgumentException - If index is not present.

setValueAtLocation

public void setValueAtLocation(int location,
                               double value)

Sets the value at the given location.

arrayCopyFrom

public final void arrayCopyFrom(double[] a)

Copy values from an array into this vector. The array should have the same size as the vector

arrayCopyFrom

public final int arrayCopyFrom(double[] a,
                               int startingArrayLocation)

Copy values from an array starting at a particular location into this vector. The array must have at least as many values beyond the starting location as there are in the vector.

Returns:

Next uncopied location in the array.

map

public final void map(java.lang.reflect.Method f)
               throws java.lang.IllegalAccessException,
                      java.lang.Throwable

Applies the method argument to each value in a non-binary vector. The method should both accept a Double as an argument and return a Double.

Throws:

java.lang.IllegalArgumentException - If the method argument has an inappropriate signature.

java.lang.UnsupportedOperationException - If vector is binary

java.lang.IllegalAccessException - If the method is inaccessible

java.lang.Throwable - If the method throws an exception it is relayed

arrayCopyInto

public final int arrayCopyInto(double[] array,
                               int startingArrayLocation)

Copy the contents of this vector into an array starting at a particular location.

Returns:

Next available location in the array

dotProduct

public double dotProduct(double[] v)

VECTOR OPERATIONS

dotProduct

public double dotProduct(ConstantMatrix m)

Specified by:

dotProduct in interface ConstantMatrix

dotProduct

public double dotProduct(DenseVector v)

extendedDotProduct

public double extendedDotProduct(DenseVector v)

dotProduct

public double dotProduct(SparseVector v)

extendedDotProduct

public double extendedDotProduct(SparseVector v)

vectorAdd

public SparseVector vectorAdd(SparseVector v,
                              double scale)

oneNorm

public double oneNorm()

Specified by:

oneNorm in interface ConstantMatrix

absNorm

public double absNorm()

Specified by:

absNorm in interface ConstantMatrix

twoNorm

public double twoNorm()

Specified by:

twoNorm in interface ConstantMatrix

infinityNorm

public double infinityNorm()

Specified by:

infinityNorm in interface ConstantMatrix

print

public void print()

Specified by:

print in interface ConstantMatrix

isNaN

public boolean isNaN()

Specified by:

isNaN in interface ConstantMatrix

isInfinite

public boolean isInfinite()

isNaNOrInfinite

public boolean isNaNOrInfinite()

sortIndices

protected void sortIndices()

removeDuplicates

protected void removeDuplicates(int numDuplicates)