jebl.evolution.trees
Class ClusteringTreeBuilder<T extends Tree>

java.lang.Object
  jebl.evolution.trees.ClusteringTreeBuilder<T>

All Implemented Interfaces:

TreeBuilder<T>

Direct Known Subclasses:

NeighborJoiningTreeBuilder

public abstract class ClusteringTreeBuilder<T extends Tree>
extends Object
implements TreeBuilder<T>

An abstract base class for clustering algorithms from pairwise distances

Author:

Andrew Rambaut, Alexei Drummond, Joseph Heled Adapted from Alexei Drummond BEAST code.

Field Summary

protected int[]	alias
protected int	besti
protected int	bestj
protected Node[]	clusters
protected double[][]	distance
protected DistanceMatrix	distanceMatrix
protected int	minimumTaxa
protected Node	newCluster
protected int	numClusters
protected int[]	tipCount

Constructor Summary

protected

ClusteringTreeBuilder(DistanceMatrix distanceMatrix, int minimumTaxa) Constructor

Method Summary

void	addProgressListener(ProgressListener listener)
T	build()
protected abstract Node	createExternalNode(Taxon taxon)
protected abstract Node	createInternalNode(Node[] nodes, double[] distances)
protected void	findNextPair() Find next two clusters to join.
protected void	finish()
void	fireSetProgress(double fractionCompleted)
static ClusteringTreeBuilder	getBuilder(TreeBuilderFactory.Method method, DistanceMatrix distances) A factory method to create a ClusteringTreeBuilder
protected double	getDist(int a, int b)
protected abstract T	getTree()
protected void	init(DistanceMatrix distanceMatrix)
protected abstract double[]	joinClusters() Inform derived class that clusters besti,bestj are being joinded into a new cluster.
protected void	newCluster()
void	removeProgressListener(ProgressListener listener)
protected abstract double	updatedDistance(int k) compute updated distance between the new cluster (besti,bestj) to any other cluster k.

Methods inherited from class java.lang.Object

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

Field Detail

distanceMatrix

protected final DistanceMatrix distanceMatrix

numClusters

protected int numClusters

clusters

protected Node[] clusters

newCluster

protected Node newCluster

besti

protected int besti

bestj

protected int bestj

tipCount

protected int[] tipCount

alias

protected int[] alias

distance

protected double[][] distance

minimumTaxa

protected int minimumTaxa

Constructor Detail

ClusteringTreeBuilder

protected ClusteringTreeBuilder(DistanceMatrix distanceMatrix,
                                int minimumTaxa)
                         throws IllegalArgumentException

Constructor

Parameters:

distanceMatrix -

minimumTaxa -

Throws:

IllegalArgumentException

Method Detail

build

public T build()

Specified by:

build in interface TreeBuilder<T extends Tree>

addProgressListener

public void addProgressListener(ProgressListener listener)

Specified by:

addProgressListener in interface TreeBuilder<T extends Tree>

removeProgressListener

public void removeProgressListener(ProgressListener listener)

Specified by:

removeProgressListener in interface TreeBuilder<T extends Tree>

fireSetProgress

public void fireSetProgress(double fractionCompleted)

getBuilder

public static ClusteringTreeBuilder getBuilder(TreeBuilderFactory.Method method,
                                               DistanceMatrix distances)

A factory method to create a ClusteringTreeBuilder

Parameters:

method - build method to use.

distances - Pre computed pairwise distances.

Returns:

A tree builder using method and distance matrix

getTree

protected abstract T getTree()

createExternalNode

protected abstract Node createExternalNode(Taxon taxon)

createInternalNode

protected abstract Node createInternalNode(Node[] nodes,
                                           double[] distances)

joinClusters

protected abstract double[] joinClusters()

Inform derived class that clusters besti,bestj are being joinded into a new cluster. New cluster will be (the smaller) besti while clusters greater than bestj are shifted one space back.

Returns:

branch distances to new internal node [besti - dist , bestj dist]

updatedDistance

protected abstract double updatedDistance(int k)

compute updated distance between the new cluster (besti,bestj) to any other cluster k. (i,j,k) are cluster indices in [0..numClusters-1]

getDist

protected double getDist(int a,
                         int b)

init

protected void init(DistanceMatrix distanceMatrix)

findNextPair

protected void findNextPair()

Find next two clusters to join. set shared best{i,j}.

newCluster

protected void newCluster()

finish

protected void finish()