weka.knowledgeflow.steps

Class Clusterer

java.lang.Object

weka.knowledgeflow.steps.BaseStep

weka.knowledgeflow.steps.WekaAlgorithmWrapper

weka.knowledgeflow.steps.Clusterer

All Implemented Interfaces:

java.io.Serializable, BaseStepExtender, PairedDataHelper.PairedProcessor<Clusterer>, Step

@KFStep(name="Clusterer",
        category="Clusterers",
        toolTipText="Weka clusterer wrapper",
        iconPath="",
        resourceIntensive=true)
public class Clusterer
extends WekaAlgorithmWrapper
implements PairedDataHelper.PairedProcessor<Clusterer>

Step that wraps a Weka clusterer. Handles trainingSet and testSet incoming connections

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected Clusterer	m_clustererTemplate The template for constructing concrete instances of the clusterer to train
protected Data	m_incrementalData Re-usable Data object for incrementalClusterer output
protected boolean	m_isReset True if the step has just been reset
protected java.io.File	m_loadModelFileName Optional file to load a pre-trained model to score with (batch, or to score and update (incremental) in the case of testSet only (batch) or instance (incremental) connections
protected boolean	m_streaming True if there is a single incoming "instance" connection
protected Clusterer	m_trainedClusterer Holds a trained/loaded clusterer
protected Instances	m_trainedClustererHeader Header used to train the clusterer
protected PairedDataHelper<Clusterer>	m_trainTestHelper Handles train test pair routing and synchronization for us

Fields inherited from class weka.knowledgeflow.steps.WekaAlgorithmWrapper

m_defaultIconPath, m_defaultPackageIconPath, m_iconPath, m_wrappedAlgorithm

Fields inherited from class weka.knowledgeflow.steps.BaseStep

m_stepIsResourceIntensive, m_stepManager, m_stepName

Constructor Summary

Constructors

Constructor and Description
Clusterer()

Method Summary

Modifier and Type	Method and Description
Clusterer	getClusterer() Get the clusterer to train
java.util.List<java.lang.String>	getIncomingConnectionTypes() Get a list of connection types that could be made to this Step at this point in time
java.io.File	getLoadClustererFileName() Get the name of the clusterer to load at execution time.
java.util.List<java.lang.String>	getOutgoingConnectionTypes() Get a list of outgoing connections that could be made from this step at this point in time
java.lang.Class	getWrappedAlgorithmClass() Get the class of the wrapped algorithm
protected void	loadModel(java.lang.String filePath) Load a pre-trained model from the supplied path
protected void	outputBatchClusterer(Clusterer clusterer, int setNum, int maxSetNum, Instances trainingSplit, Instances testSplit) Output batch clusterer data to downstream steps
protected void	outputGraphData(Clusterer clusterer, int setNum) Output a Data object containing a dot graph, if the model is Drawable and we have downstream steps receiving graph connections.
protected void	outputTextData(Clusterer clusterer, int setNum) Output a Data object containing a textual description of a model to any outgoing text connections
void	processIncoming(Data data) Process an incoming data object
protected void	processOnlyTestSet(Data data) Process a Data object in the case where we only have an incoming testSet connection
Clusterer	processPrimary(java.lang.Integer setNum, java.lang.Integer maxSetNum, Data data, PairedDataHelper<Clusterer> helper) Process a training split (primary data handled by the PairedDataHelper)
void	processSecondary(java.lang.Integer setNum, java.lang.Integer maxSetNum, Data data, PairedDataHelper<Clusterer> helper) Process a test split/fold (secondary data handled by PairedDataHelper)
void	setClusterer(Clusterer clusterer) Set the clusterer to train
void	setLoadClustererFileName(java.io.File filename) Set the name of the clusterer to load at execution time.
void	setWrappedAlgorithm(java.lang.Object algo) Set the wrapped algorithm
void	stepInit() Initialize the step

Methods inherited from class weka.knowledgeflow.steps.WekaAlgorithmWrapper

getDefaultIconPath, getDefaultPackageLevelIconPath, getIconPath, getWrappedAlgorithm, globalInfo

Methods inherited from class weka.knowledgeflow.steps.BaseStep

environmentSubstitute, getCustomEditorForStep, getDefaultSettings, getInteractiveViewers, getInteractiveViewersImpls, getName, getStepManager, isResourceIntensive, isStopRequested, outputStructureForConnectionType, setName, setStepIsResourceIntensive, setStepManager, setStepMustRunSingleThreaded, start, stepMustRunSingleThreaded, stop

Methods inherited from class java.lang.Object

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

Field Detail

m_clustererTemplate

protected Clusterer m_clustererTemplate

The template for constructing concrete instances of the clusterer to train

m_trainedClusterer

protected Clusterer m_trainedClusterer

Holds a trained/loaded clusterer

m_trainedClustererHeader

protected Instances m_trainedClustererHeader

Header used to train the clusterer

m_trainTestHelper

protected transient PairedDataHelper<Clusterer> m_trainTestHelper

Handles train test pair routing and synchronization for us

m_loadModelFileName

protected java.io.File m_loadModelFileName

Optional file to load a pre-trained model to score with (batch, or to score and update (incremental) in the case of testSet only (batch) or instance (incremental) connections

m_isReset

protected boolean m_isReset

True if the step has just been reset

m_incrementalData

protected Data m_incrementalData

Re-usable Data object for incrementalClusterer output

m_streaming

protected boolean m_streaming

True if there is a single incoming "instance" connection

Constructor Detail

Clusterer

public Clusterer()

Method Detail

getClusterer

public Clusterer getClusterer()

Get the clusterer to train

Returns:

the clusterer to train

setClusterer

@ProgrammaticProperty
public void setClusterer(Clusterer clusterer)

Set the clusterer to train

Parameters:

clusterer - the clusterer to train

getLoadClustererFileName

public java.io.File getLoadClustererFileName()

Get the name of the clusterer to load at execution time. This only applies in the case where the only incoming connection is a test set connection (batch mode) or an instance connection (incremental prediction mode).

Returns:

the name of the file to load the model from

setLoadClustererFileName

@OptionMetadata(displayName="Clusterer model to load",
                description="Optional path to a clusterer to load at execution time (only applies when using testSet connections)")
 @FilePropertyMetadata(fileChooserDialogType=0,
                      directoriesOnly=false)
public void setLoadClustererFileName(java.io.File filename)

Set the name of the clusterer to load at execution time. This only applies in the case where the only incoming connection is a test set connection (batch mode) or an instance connection (incremental prediction mode).

Parameters:

filename - the name of the file to load the model from

getWrappedAlgorithmClass

public java.lang.Class getWrappedAlgorithmClass()

Get the class of the wrapped algorithm

Specified by:

getWrappedAlgorithmClass in class WekaAlgorithmWrapper

Returns:

the class of the wrapped algorithm

setWrappedAlgorithm

public void setWrappedAlgorithm(java.lang.Object algo)

Set the wrapped algorithm

Overrides:

setWrappedAlgorithm in class WekaAlgorithmWrapper

Parameters:

algo - the algorithm to wrap

stepInit

public void stepInit()
              throws WekaException

Initialize the step

Specified by:

stepInit in interface BaseStepExtender

Specified by:

stepInit in interface Step

Throws:

WekaException - if a problem occurs

processIncoming

public void processIncoming(Data data)
                     throws WekaException

Process an incoming data object

Specified by:

processIncoming in interface BaseStepExtender

Specified by:

processIncoming in interface Step

Overrides:

processIncoming in class BaseStep

Parameters:

data - the data to process

Throws:

WekaException - if a problem occurs

processOnlyTestSet

protected void processOnlyTestSet(Data data)
                           throws WekaException

Process a Data object in the case where we only have an incoming testSet connection

Parameters:

data - the Data object to process

Throws:

WekaException - if a problem occurs

getIncomingConnectionTypes

public java.util.List<java.lang.String> getIncomingConnectionTypes()

Get a list of connection types that could be made to this Step at this point in time

Specified by:

getIncomingConnectionTypes in interface BaseStepExtender

Specified by:

getIncomingConnectionTypes in interface Step

Returns:

a list of incoming connection types that could be made at this time

getOutgoingConnectionTypes

public java.util.List<java.lang.String> getOutgoingConnectionTypes()

Get a list of outgoing connections that could be made from this step at this point in time

Specified by:

getOutgoingConnectionTypes in interface BaseStepExtender

Specified by:

getOutgoingConnectionTypes in interface Step

Returns:

a list of outgoing connections that could be made at this point in time

loadModel

protected void loadModel(java.lang.String filePath)
                  throws java.lang.Exception

Load a pre-trained model from the supplied path

Parameters:

filePath - the path to load the model from

Throws:

java.lang.Exception - if a problem occurs

outputGraphData

protected void outputGraphData(Clusterer clusterer,
                               int setNum)
                        throws WekaException

Output a Data object containing a dot graph, if the model is Drawable and we have downstream steps receiving graph connections.

Parameters:

clusterer - the clusterer to generate the graph from

setNum - the set number of the data used to generate the graph

Throws:

WekaException - if a problem occurs

outputTextData

protected void outputTextData(Clusterer clusterer,
                              int setNum)
                       throws WekaException

Output a Data object containing a textual description of a model to any outgoing text connections

Parameters:

clusterer - the clusterer to get the textual description of

setNum - the set number of the training data

Throws:

WekaException - if a problem occurs

outputBatchClusterer

protected void outputBatchClusterer(Clusterer clusterer,
                                    int setNum,
                                    int maxSetNum,
                                    Instances trainingSplit,
                                    Instances testSplit)
                             throws WekaException

Output batch clusterer data to downstream steps

Parameters:

clusterer - the clusterer to outpit

setNum - the set number of the current dataset

maxSetNum - the maximum set number

trainingSplit - the training data

testSplit - the test data, or null if there is no test data

Throws:

WekaException - if a problem occurs

processPrimary

public Clusterer processPrimary(java.lang.Integer setNum,
                                java.lang.Integer maxSetNum,
                                Data data,
                                PairedDataHelper<Clusterer> helper)
                         throws WekaException

Process a training split (primary data handled by the PairedDataHelper)

Specified by:

processPrimary in interface PairedDataHelper.PairedProcessor<Clusterer>

Parameters:

setNum - the number of this split/fold

maxSetNum - the maximum number of splits/folds in the group

data - the actual split/fold data

helper - the PairedDataHelper managing the paired data

Returns:

a Classifier trained on this training split

Throws:

WekaException - if a problem occurs

processSecondary

public void processSecondary(java.lang.Integer setNum,
                             java.lang.Integer maxSetNum,
                             Data data,
                             PairedDataHelper<Clusterer> helper)
                      throws WekaException

Process a test split/fold (secondary data handled by PairedDataHelper)

Specified by:

processSecondary in interface PairedDataHelper.PairedProcessor<Clusterer>

Parameters:

setNum - the set number of this split/fold

maxSetNum - the maximum number of splits/folds in the group

data - the actual split/fold data

helper - the PairedDataHelper managing the paried data

Throws:

WekaException - if a problem occurs