org.ai_heuristic.algo.cluster

Class NearestNeighbour

java.lang.Object

org.ai_heuristic.functs.Function

org.ai_heuristic.algo.AlgorithmCompare

org.ai_heuristic.algo.AlgorithmCompareTree

org.ai_heuristic.algo.cluster.NearestNeighbour

All Implemented Interfaces:

FunctionDef

public class NearestNeighbour
extends AlgorithmCompareTree

This implement a version of the nearest neighbour algorithm that searches over a balanced KD_Tree tree.

Author:

Kieran Greer

Field Summary

Fields

Modifier and Type	Field and Description
protected int	ksize The number of nearest neighbours to return
protected EvaluateMathDef	mathCompare Single value math comparisons - simple type

Fields inherited from class org.ai_heuristic.algo.AlgorithmCompareTree

bs, kdTree

Fields inherited from class org.ai_heuristic.algo.AlgorithmCompare

bestEval, evalMetric

Fields inherited from class org.ai_heuristic.functs.Function

config, valueType

Constructor Summary

Constructors

Constructor and Description
NearestNeighbour(FunctionMetric thisEvalMetric, EvaluateMathDef mathEval) Create a new empty instance of NearestNeighbour.
NearestNeighbour(java.lang.String thisValueType, java.util.HashMap<java.lang.String,?> thisConfig) Create a new instance of NearestNeighbour.

Method Summary

Modifier and Type	Method and Description
ReplySet	evaluate(MetricDataset cDataset) Calculate the closest ksize data points from each pont to the other and use this to cluster.
protected void	initialise() Initialise the function values, setting the config parameters or other.
protected java.util.ArrayList<java.util.ArrayList<java.lang.String>>	toClusters(java.util.HashMap<java.lang.String,java.util.ArrayList<KD_TreeNode>> nnGroups) Convert the groups of nn nodes to clusters of node names.

Methods inherited from class org.ai_heuristic.algo.AlgorithmCompareTree

getDatasetTree, resetValues, setDatasetTree, updateTree

Methods inherited from class org.ai_heuristic.algo.AlgorithmCompare

getBestEvaluation

Methods inherited from class org.ai_heuristic.functs.Function

checkValueType, createFunction, createFunction, createFunction, getConfigParams, innerObject, isLegalNumber, setConfigParams, setEvaluator, setValueType

Methods inherited from class java.lang.Object

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

Field Detail

ksize

protected int ksize

The number of nearest neighbours to return

mathCompare

protected EvaluateMathDef mathCompare

Single value math comparisons - simple type

Constructor Detail

NearestNeighbour

public NearestNeighbour(FunctionMetric thisEvalMetric,
                        EvaluateMathDef mathEval)
                 throws java.lang.Exception

Create a new empty instance of NearestNeighbour. Set the local dataset later.

Parameters:

thisEvalMetric - the distance function.

mathEval - for simple math comparisons.

Throws:

java.lang.Exception - any error.

NearestNeighbour

public NearestNeighbour(java.lang.String thisValueType,
                        java.util.HashMap<java.lang.String,?> thisConfig)
                 throws java.lang.Exception

Create a new instance of NearestNeighbour.

Parameters:

thisValueType - the type of object being evaluated. Can be null if set later or not used.

thisConfig - list of initialisation function-specific parameters.

Use AiHeuristicConst.KNN to store the number of nearest neighbours value, .DIM to store the number of dimensions value, .EVALUATOR to store the instance of the base mathCompare evaluator.

Throws:

java.lang.Exception - any error.

Method Detail

initialise

protected void initialise()

Initialise the function values, setting the config parameters or other.

Overrides:

initialise in class AlgorithmCompareTree

evaluate

public ReplySet evaluate(MetricDataset cDataset)
                  throws java.lang.Exception

Calculate the closest ksize data points from each pont to the other and use this to cluster.

Specified by:

evaluate in interface FunctionDef

Overrides:

evaluate in class Function

Parameters:

cDataset - the dataset to cluster, as specified above. Add all points to the tree and then return clusters for the new dataset points.

Returns:

the cluster indexes from the datasets list, as ArrayList lists. You can use ((Integer)ReplySet.getValue()).intValue() to retrieve the cluster values.

Throws:

java.lang.Exception - any error.

toClusters

protected java.util.ArrayList<java.util.ArrayList<java.lang.String>> toClusters(java.util.HashMap<java.lang.String,java.util.ArrayList<KD_TreeNode>> nnGroups)

Convert the groups of nn nodes to clusters of node names.

Parameters:

nnGroups - the groups of nn nodes to cluster.

Returns:

the clustered names lists.