boone.map

Class HexagonTopology

java.lang.Object

boone.map.Topology

boone.map.HexagonTopology

All Implemented Interfaces:

Storable, java.lang.Cloneable

public class HexagonTopology
extends Topology

A hexagonal output map.

Author:

Helmut A. Mayer

Field Summary

Fields inherited from class boone.map.Topology

EMPTY, net

Constructor Summary

Constructors

Constructor and Description
HexagonTopology(int rows, int columns) Creates the SOM Trainer.

Method Summary

Modifier and Type	Method and Description
boolean	arrangeNeurons() Arranges the output neurons of the net according to the topology, i.e., the positions of neurons are set and the neurons get an ID, which can be used to identify its position in the map.
protected void	clearMap() Clears the map (set to EMPTY).
HexagonTopology	clone() A deep clone.
void	fromXML(org.jdom2.Element element) Reads attributes from XML tree.
Neuron	getLowerLeft(int id) Returns the lower left neighbor neuron, if there.
Neuron	getLowerRight(int id) Returns the lower right neighbor neuron, if there.
int[]	getNeurons() Returns the dimensions of the map.
Neuron	getRight(int id) Returns the right neighbor neuron, if there.
double[]	getSizes() Returns the size (length) of each dimension of the map.
int	getTotalNeurons() Returns the total number of map neurons.
protected void	searchNeighbors(int id, int radius) Searches direct hexagonal neighbors of the given center neuron.
protected void	setMapValue(int index, int value) Sets a map value at the spot corresponding to the neuron index.
org.jdom2.Element	toXML(org.jdom2.Element element) Writes specific attributes to XML tree.

Methods inherited from class boone.map.Topology

addNeighbor, getNeighbors, getNet, setNet

Methods inherited from class java.lang.Object

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

Constructor Detail

HexagonTopology

public HexagonTopology(int rows,
                       int columns)

Creates the SOM Trainer. At this time the associated net is not known. Could be improved, but maybe too complicated.

Method Detail

clone

public HexagonTopology clone()

A deep clone.

Overrides:

clone in class Topology

Returns:

a clone, null if problems which are very unlikely

getNeurons

public int[] getNeurons()

Returns the dimensions of the map.

Specified by:

getNeurons in class Topology

Returns:

the number of rows and columns

getSizes

public double[] getSizes()

Returns the size (length) of each dimension of the map.

Specified by:

getSizes in class Topology

Returns:

the length of each dimension

arrangeNeurons

public boolean arrangeNeurons()

Arranges the output neurons of the net according to the topology, i.e., the positions of neurons are set and the neurons get an ID, which can be used to identify its position in the map. The horizontal distances of neurons are 1. Each second row (starting with row 1) is offset with 0.5 to the right, which reflects the hexagonal structure. Note that for neighborhood calculations the distance of neighboring neurons is always 1, i.e., the six hexagonal neighbors of a neuron all have distance 1.

Specified by:

arrangeNeurons in class Topology

Returns:

true, if arranged, false, if the number of output neurons does not match the map size

setMapValue

protected void setMapValue(int index,
                           int value)

Sets a map value at the spot corresponding to the neuron index.

Specified by:

setMapValue in class Topology

Parameters:

index - a neuron index

value - the map value

getTotalNeurons

public int getTotalNeurons()

Returns the total number of map neurons.

Specified by:

getTotalNeurons in class Topology

Returns:

the total number of neurons in the map

clearMap

protected void clearMap()

Clears the map (set to EMPTY).

Specified by:

clearMap in class Topology

searchNeighbors

protected void searchNeighbors(int id,
                               int radius)

Searches direct hexagonal neighbors of the given center neuron. The distance of the neighbors of the start neuron is entered in the map at the corresponding neuron position. The start neuron has a distance of 0. The first neighbors a distance of 1, etc. All unchecked neuron map positions have a distance of EMPTY(-1). If the center neuron has a distance of the given radius, it is a border neuron, so search is stopped.

Specified by:

searchNeighbors in class Topology

Parameters:

id - the center neuron index

radius - the neighborhood radius

fromXML

public void fromXML(org.jdom2.Element element)

Description copied from interface: Storable

Reads attributes from XML tree.

Parameters:

element - a JDOM element

toXML

public org.jdom2.Element toXML(org.jdom2.Element element)

Writes specific attributes to XML tree.

Specified by:

toXML in interface Storable

Overrides:

toXML in class Topology

Parameters:

element - the parent element

Returns:

the topology element

getRight

public Neuron getRight(int id)

Returns the right neighbor neuron, if there.

Parameters:

id - the neuron ID

Returns:

the right neighbor, null if none

getLowerLeft

public Neuron getLowerLeft(int id)

Returns the lower left neighbor neuron, if there.

Parameters:

id - the neuron ID

Returns:

the lower left neighbor, null if none

getLowerRight

public Neuron getLowerRight(int id)

Returns the lower right neighbor neuron, if there.

Parameters:

id - the neuron ID

Returns:

the lower right neighbor, null if none