org.knowceans.mcl
Class SparseVector

java.lang.Object
  java.util.AbstractMap<K,V>
      java.util.HashMap<java.lang.Integer,java.lang.Double>
          org.knowceans.mcl.SparseVector

All Implemented Interfaces:

java.io.Serializable, java.lang.Cloneable, java.util.Map<java.lang.Integer,java.lang.Double>

public class SparseVector
extends java.util.HashMap<java.lang.Integer,java.lang.Double>

SparseVector represents a sparse vector.

Conventions: except for the inherited methods and normalise(double), operations leave this ummodified (immutable) if there is a return value. Within operations, no pruning of values close to zero is done. Pruning can be controlled via the prune() method.

Author:

gregor :: arbylon . net

See Also:

Serialized Form

Field Summary

private int	length
private static long	serialVersionUID

Constructor Summary

SparseVector()
create empty vector

SparseVector(double[] x)
create vector from dense vector

SparseVector(int i)
create empty vector with length

SparseVector(SparseVector v)
copy constructor

Method Summary

void	add(int i, double a) mutable add
void	add(SparseVector v) mutable add
SparseVector	copy() copy the contents of the sparse vector
double	expSum(int p) exponential sum, i.e., sum (elements^p)
void	factor(double a) mutable factorisation
java.lang.Double	get(java.lang.Object key) get ensures it returns 0 for empty hash values or if index exceeds length.
double[]	getDense() get dense represenation
int	getLength() get the length of the vector
void	hadamardPower(double s) mutable Hadamard power
void	hadamardProduct(SparseVector v) mutable Hadamard product (elementwise multiplication)
double	max() maximum element value
void	mult(int i, double a) mutable mult
void	normalise() normalises the vector to 1.
double	normalise(double newsum) normalises the vector to newsum
void	prune(double threshold) remove all elements whose magnitude is < threshold
java.lang.Double	put(java.lang.Integer key, java.lang.Double value) put increases the matrix size if the index exceeds the current size.
void	setLength(int length) set the new length of the vector (regardless of the maximum index).
private double	sum() sum of the elements
double	sum(double s) power sum of the elements
double	times(SparseVector v) immutable scalar product
java.lang.String	toString()
java.lang.String	toStringDense() create string representation of dense equivalent.

Methods inherited from class java.util.HashMap

clear, clone, containsKey, containsValue, entrySet, isEmpty, keySet, putAll, remove, size, values

Methods inherited from class java.util.AbstractMap

equals, hashCode

Methods inherited from class java.lang.Object

finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface java.util.Map

equals, hashCode

Field Detail

serialVersionUID

private static final long serialVersionUID

See Also:

Constant Field Values

length

private int length

Constructor Detail

SparseVector

public SparseVector()

create empty vector

SparseVector

public SparseVector(int i)

create empty vector with length

SparseVector

public SparseVector(double[] x)

create vector from dense vector

Parameters:

x -

SparseVector

public SparseVector(SparseVector v)

copy constructor

Parameters:

v -

Method Detail

get

public java.lang.Double get(java.lang.Object key)

get ensures it returns 0 for empty hash values or if index exceeds length.

Specified by:

get in interface java.util.Map<java.lang.Integer,java.lang.Double>

Overrides:

get in class java.util.HashMap<java.lang.Integer,java.lang.Double>

Parameters:

key -

Returns:

val

put

public java.lang.Double put(java.lang.Integer key,
                            java.lang.Double value)

put increases the matrix size if the index exceeds the current size.

Specified by:

put in interface java.util.Map<java.lang.Integer,java.lang.Double>

Overrides:

put in class java.util.HashMap<java.lang.Integer,java.lang.Double>

Parameters:

key -

value -

Returns:

normalise

public void normalise()

normalises the vector to 1.

normalise

public double normalise(double newsum)

normalises the vector to newsum

Parameters:

the - value to which the element sum

Returns:

the old element sum

sum

private double sum()

sum of the elements

Returns:

sum

public double sum(double s)

power sum of the elements

Returns:

add

public void add(SparseVector v)

mutable add

Parameters:

v -

mult

public void mult(int i,
                 double a)

mutable mult

Parameters:

i - index

a - value

factor

public void factor(double a)

mutable factorisation

Parameters:

a -

times

public double times(SparseVector v)

immutable scalar product

Parameters:

v -

Returns:

scalar product

hadamardProduct

public void hadamardProduct(SparseVector v)

mutable Hadamard product (elementwise multiplication)

Parameters:

v -

hadamardPower

public void hadamardPower(double s)

mutable Hadamard power

Parameters:

s -

add

public void add(int i,
                double a)

mutable add

Parameters:

i -

a -

getLength

public final int getLength()

get the length of the vector

Returns:

setLength

public final void setLength(int length)

set the new length of the vector (regardless of the maximum index).

Parameters:

length -

copy

public SparseVector copy()

copy the contents of the sparse vector

Returns:

toString

public java.lang.String toString()

Overrides:

toString in class java.util.AbstractMap<java.lang.Integer,java.lang.Double>

toStringDense

public java.lang.String toStringDense()

create string representation of dense equivalent.

Returns:

getDense

public double[] getDense()

get dense represenation

Returns:

max

public double max()

maximum element value

Returns:

expSum

public double expSum(int p)

exponential sum, i.e., sum (elements^p)

Parameters:

p -

Returns:

prune

public void prune(double threshold)

remove all elements whose magnitude is < threshold

Parameters:

threshold -