it.uniroma2.sag.kelp.data.representation.vector

Class SparseVector

java.lang.Object

it.uniroma2.sag.kelp.data.representation.vector.SparseVector

All Implemented Interfaces:

Normalizable, Representation, Vector, Serializable

public class SparseVector
extends Object
implements Vector

Sparse Feature Vector. It is based on a Hashmap implementation.

Author:

Simone Filice

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
static float	threshold

Constructor Summary

Constructors

Constructor and Description
SparseVector()

Method Summary

Methods

Modifier and Type	Method and Description
void	add(float coeff, float vectorCoeff, Vector vector) Add a vector multiplied by vectorCoeff to this vector multiplied by
void	add(float coeff, Vector vector) Add a vector multiplied by coeff to this vector
void	add(Vector vector) Add a vector to this vector
SparseVector	copyVector() Returns a copy of this vector.
float	euclideanDistance(Vector vector) Returns the euclidean distance between this vector and vector
Map<Object,Number>	getActiveFeatures() Returns a map containing all the non-zero features
float	getFeatureValue(Object featureIdentifier) Returns the value of the feature identified with featureIdentifier
float	getFeatureValue(String featureName) Returns the value associated to a feature
static gnu.trove.map.TIntObjectMap<String>	getFromIntToWord()
static gnu.trove.map.TObjectIntMap<String>	getFromWordToInt()
float	getSquaredNorm() Returns the squared norm of this vector
String	getTextFromData() Returns a textual representation of the data stored in this representation
gnu.trove.map.TIntFloatMap	getVector()
static int	getWordCounter()
SparseVector	getZeroVector() Returns a vector whose values are all 0.
void	incrementFeature(String featureName, float valueIncrement) Increments the value associated to a feature
float	innerProduct(Vector vector) Returns the dot product between this vector and vector
boolean	isCompatible(Representation rep) Evaluates whether a representation is compatible with this one, e.g.., they have same type and pass additional checks that are type-dependent
void	merge(Vector vector, float coefficient, String newDimensionPrefix) Merge this vector with vector (it is like a vector concatenation) If V1 is the space where this vector lies and V2 is the space where vector lies, then the resulting vector lies in V1xV2
void	normalize() Scales the representation in order to have a unit norm in the explicit feature space
void	pointWiseProduct(Vector vector) Compute the point-wise product of this vector with the one in vector.
void	scale(float coeff) Multiplies each element of this representation by coeff
void	setDataFromText(String representationDescription) Initializes a Representation using its textual description provided in representationDescription
void	setFeatureValue(Object featureIdentifier, float value) Assigns value to the feature identified by featureIdentifier
void	setFeatureValue(String featureName, float value) Sets the value of a feature
void	setVector(gnu.trove.map.TIntFloatMap vector)
static void	setWordCounter(int wordCounter)
String	toString()

Methods inherited from class java.lang.Object

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

Field Detail

threshold

public static final float threshold

See Also:

Constant Field Values

Constructor Detail

SparseVector

public SparseVector()

Method Detail

normalize

public void normalize()

Description copied from interface: Normalizable

Scales the representation in order to have a unit norm in the explicit feature space

Specified by:

normalize in interface Normalizable

setDataFromText

public void setDataFromText(String representationDescription)
                     throws IOException

Description copied from interface: Representation

Initializes a Representation using its textual description provided in representationDescription

Specified by:

setDataFromText in interface Representation

Parameters:

representationDescription - the textual description of the representation to be initialized

Throws:

IOException

toString

public String toString()

Overrides:

toString in class Object

getVector

public gnu.trove.map.TIntFloatMap getVector()

Returns:

the vector

setVector

public void setVector(gnu.trove.map.TIntFloatMap vector)

Parameters:

vector - the vector to set

innerProduct

public float innerProduct(Vector vector)

Description copied from interface: Vector

Returns the dot product between this vector and vector

Specified by:

innerProduct in interface Vector

Returns:

the dot product

scale

public void scale(float coeff)

Description copied from interface: Normalizable

Multiplies each element of this representation by coeff

Specified by:

scale in interface Normalizable

Parameters:

coeff - the scaling factor

add

public void add(Vector vector)

Description copied from interface: Vector

Add a vector to this vector

Specified by:

add in interface Vector

Parameters:

vector - the vector to be added

add

public void add(float coeff,
       Vector vector)

Description copied from interface: Vector

Add a vector multiplied by coeff to this vector

Specified by:

add in interface Vector

Parameters:

coeff - the vector coefficient

vector - the vector to be added

add

public void add(float coeff,
       float vectorCoeff,
       Vector vector)

Description copied from interface: Vector

Add a vector multiplied by vectorCoeff to this vector multiplied by

Specified by:

add in interface Vector

Parameters:

coeff - the coefficient of this vector

vectorCoeff - the vector coefficient

vector - the vector to be added

getZeroVector

public SparseVector getZeroVector()

Description copied from interface: Vector

Returns a vector whose values are all 0. The returned vector has the same dimensions of this

Specified by:

getZeroVector in interface Vector

Returns:

a zero vector

getTextFromData

public String getTextFromData()

Description copied from interface: Representation

Returns a textual representation of the data stored in this representation

Specified by:

getTextFromData in interface Representation

Returns:

a textual representation of the data stored in this representation

getSquaredNorm

public float getSquaredNorm()

Description copied from interface: Normalizable

Returns the squared norm of this vector

Specified by:

getSquaredNorm in interface Normalizable

Returns:

the squared norm

getActiveFeatures

public Map<Object,Number> getActiveFeatures()

Description copied from interface: Vector

Returns a map containing all the non-zero features

Specified by:

getActiveFeatures in interface Vector

Returns:

the non zero features

merge

public void merge(Vector vector,
         float coefficient,
         String newDimensionPrefix)

Merge this vector with vector (it is like a vector concatenation) If V1 is the space where this vector lies and V2 is the space where vector lies, then the resulting vector lies in V1xV2

NOTE: this is not a sum because even if some feature names are shared between the two vectors, those will be considered different dimensions (a prefix is added to all feature names of vector)

Parameters:

vector - the input vector to be merged with this

coefficient - a scaling factor for vector

newDimensionPrefix - the prefix to be added to all the feature names of vector during the merging process (vector is not modified)

pointWiseProduct

public void pointWiseProduct(Vector vector)

Description copied from interface: Vector

Compute the point-wise product of this vector with the one in vector.

Specified by:

pointWiseProduct in interface Vector

Parameters:

vector - the vector used for the point-wise product

copyVector

public SparseVector copyVector()

Description copied from interface: Vector

Returns a copy of this vector.

Specified by:

copyVector in interface Vector

Returns:

Vector

setFeatureValue

public void setFeatureValue(String featureName,
                   float value)

Sets the value of a feature

Parameters:

featureName - the name of the feature

value - the value of the feature

getFeatureValue

public float getFeatureValue(String featureName)

Returns the value associated to a feature

Parameters:

featureName - the identifier of the feature

Returns:

the feature value

incrementFeature

public void incrementFeature(String featureName,
                    float valueIncrement)

Increments the value associated to a feature

Parameters:

featureName - the identifier of the feature

valueIncrement - the increment

setFeatureValue

public void setFeatureValue(Object featureIdentifier,
                   float value)

Description copied from interface: Vector

Assigns value to the feature identified by featureIdentifier

NOTE: this method could be not the most efficient to set a feature value. Actual implementations of the Vector class may provide faster methods using their specific type to identify features (instead of a generic Object)

Specified by:

setFeatureValue in interface Vector

Parameters:

featureIdentifier - the identifier of the feature

value - the value of the feature

getFeatureValue

public float getFeatureValue(Object featureIdentifier)

Description copied from interface: Vector

Returns the value of the feature identified with featureIdentifier

NOTE: this method could be not the most efficient to get a feature value. Actual implementations of the Vector class may provide faster methods using their specific type to identify features (instead of a generic Object)

Specified by:

getFeatureValue in interface Vector

Returns:

the value of the feature

euclideanDistance

public float euclideanDistance(Vector vector)
                        throws VectorOperationException

Description copied from interface: Vector

Returns the euclidean distance between this vector and vector

Specified by:

euclideanDistance in interface Vector

Returns:

the euclidead distance

Throws:

VectorOperationException

isCompatible

public boolean isCompatible(Representation rep)

Description copied from interface: Representation

Evaluates whether a representation is compatible with this one, e.g.., they have same type and pass additional checks that are type-dependent

Specified by:

isCompatible in interface Representation

Parameters:

rep - the representation to be compared with this one

Returns:

whether this representation is compatible with the input one

setWordCounter

public static void setWordCounter(int wordCounter)

getFromIntToWord

public static gnu.trove.map.TIntObjectMap<String> getFromIntToWord()

getFromWordToInt

public static gnu.trove.map.TObjectIntMap<String> getFromWordToInt()

getWordCounter

public static int getWordCounter()