adams.flow.webservice
Class SimpleWekaService

java.lang.Object
  adams.core.ConsoleObject
      adams.core.option.AbstractOptionHandler
          adams.flow.webservice.SimpleWekaService

All Implemented Interfaces:

Debuggable, Destroyable, OptionHandler, SizeOfHandler, OwnedByWekaServiceWS, Serializable, WekaService

public class SimpleWekaService
extends AbstractOptionHandler
implements WekaService, OwnedByWekaServiceWS

Class that implements the weka web service.

Version:

$Revision$

Author:

msf8, FracPete (fracpete at waikato dot ac dot nz)

See Also:

Serialized Form

Field Summary

protected LRUCache<String,Classifier>	m_Classifiers the cache for the classifier models.
protected LRUCache<String,weka.clusterers.Clusterer>	m_Clusterers the cache for the clusterer models.
protected WekaServiceWS	m_Owner web service object

Fields inherited from class adams.core.option.AbstractOptionHandler

m_DebugLevel, m_OptionManager

Constructor Summary

SimpleWekaService()
Default Constructor.

Method Summary

CrossValidateResponseObject	crossValidateClassifier(Dataset dataset, int seed, int folds, String classifier) Cross-validates a classifier on a dataset.
DisplayClassifierResponseObject	displayClassifier(String model) Returns the string representation of the specified classifier.
DisplayClustererResponseObject	displayClusterer(String model) Returns the string representation of the specified clusterer.
protected void	displayString(Dataset dataset) For outputting the dataset in debug mode.
WekaServiceWS	getOwner() Returns the current owner of this webservice.
String	globalInfo() Returns a string for the GUI that describes this object.
List<String>	listClassifiers() Returns a list of all classifier models currently stored on the server.
List<String>	listClusterers() Returns a list of all clusterer models currently stored on the server.
OptimizeReturnObject	optimizeClassifierMultiSearch(String classifier, List<String> searchParameters, Dataset dataset, String evaluation) Optimizes the parameters of a base classifier using MultiSearch.
PredictClassifierResponseObject	predictClassifier(Dataset dataset, String modelName) Makes predictions using a previously generated model.
PredictClustererResponseObject	predictClusterer(Dataset dataset, String modelName) Uses a previously built clusterer model to predict clusters for the provided data.
void	setOwner(WekaServiceWS value) Sets the owner of this webservice.
TestClassifierResponseObject	testClassifier(Dataset dataset, String modelName) Tests a previously trained model.
TrainClassifierResponseObject	trainClassifier(Dataset dataset, String classifier, String name) Performs training of a classifier and stores it in the model cache.
TrainClustererResponseObject	trainClusterer(Dataset dataset, String clusterer, String modelName) Trains a cluster algorithm.
TransformResponseObject	transform(Dataset dataset, String actorName) Transforms a dataset using a global actor on the server.

Methods inherited from class adams.core.option.AbstractOptionHandler

cleanUpOptions, debug, debug, debugLevelTipText, defineOptions, destroy, finishInit, getDebugLevel, getOptionManager, initialize, isDebugOn, newOptionManager, reset, setDebugLevel, toCommandLine, toString

Methods inherited from class adams.core.ConsoleObject

getDebugging, getSystemErr, getSystemOut, sizeOf

Methods inherited from class java.lang.Object

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

Field Detail

m_Owner

protected WekaServiceWS m_Owner

web service object

m_Classifiers

protected LRUCache<String,Classifier> m_Classifiers

the cache for the classifier models.

m_Clusterers

protected LRUCache<String,weka.clusterers.Clusterer> m_Clusterers

the cache for the clusterer models.

Constructor Detail

SimpleWekaService

public SimpleWekaService()

Default Constructor.

NB: the owning webservice needs to get set before using this implemention, using the setOwner(WekaServiceWS) method.

Method Detail

globalInfo

public String globalInfo()

Returns a string for the GUI that describes this object.

Specified by:

globalInfo in class AbstractOptionHandler

Returns:

the description

setOwner

public void setOwner(WekaServiceWS value)

Sets the owner of this webservice.

Specified by:

setOwner in interface OwnedByWekaServiceWS

Parameters:

value - the owner

getOwner

public WekaServiceWS getOwner()

Returns the current owner of this webservice.

Specified by:

getOwner in interface OwnedByWekaServiceWS

Returns:

the owner, null if none set

trainClassifier

public TrainClassifierResponseObject trainClassifier(Dataset dataset,
                                                     String classifier,
                                                     String name)

Performs training of a classifier and stores it in the model cache.

Specified by:

trainClassifier in interface WekaService

Parameters:

dataset - the dataset to use for training

classifier - the classifier setup

name - the identifier for the model

Returns:

null if OK, otherwise error message

testClassifier

public TestClassifierResponseObject testClassifier(Dataset dataset,
                                                   String modelName)

Tests a previously trained model.

Specified by:

testClassifier in interface WekaService

Parameters:

dataset - the dataset to use for testing

modelName - the name of the model to use

Returns:

the response

crossValidateClassifier

public CrossValidateResponseObject crossValidateClassifier(Dataset dataset,
                                                           int seed,
                                                           int folds,
                                                           String classifier)

Cross-validates a classifier on a dataset.

Specified by:

crossValidateClassifier in interface WekaService

Parameters:

dataset - the dataset to use for cross-validation

seed - the seed for randomizing the data

folds - the number of folds to use

classifier - the classifier setup

Returns:

the response

predictClassifier

public PredictClassifierResponseObject predictClassifier(Dataset dataset,
                                                         String modelName)

Makes predictions using a previously generated model.

Specified by:

predictClassifier in interface WekaService

Parameters:

dataset - the data to use for the predictions

modelName - the model to use

Returns:

the response

transform

public TransformResponseObject transform(Dataset dataset,
                                         String actorName)

Transforms a dataset using a global actor on the server.

Specified by:

transform in interface WekaService

Parameters:

dataset - the data to transform

actorName - the global actor to use

Returns:

the response with the transformed data or an error message

trainClusterer

public TrainClustererResponseObject trainClusterer(Dataset dataset,
                                                   String clusterer,
                                                   String modelName)

Trains a cluster algorithm.

Specified by:

trainClusterer in interface WekaService

Parameters:

dataset - the data to use

clusterer - the clusterer setup to use

modelName - the name to store the model under

Returns:

the response

predictClusterer

public PredictClustererResponseObject predictClusterer(Dataset dataset,
                                                       String modelName)

Uses a previously built clusterer model to predict clusters for the provided data.

Specified by:

predictClusterer in interface WekaService

Parameters:

dataset - the data to predict the clusters for

modelName - the name of the clusterer to use

Returns:

the predictions

displayClassifier

public DisplayClassifierResponseObject displayClassifier(String model)

Returns the string representation of the specified classifier.

Specified by:

displayClassifier in interface WekaService

Parameters:

model - the model name to return the string representation for

Returns:

the response

displayClusterer

public DisplayClustererResponseObject displayClusterer(String model)

Returns the string representation of the specified clusterer.

Specified by:

displayClusterer in interface WekaService

Parameters:

model - the model name to return the string representation for

Returns:

the response

listClassifiers

public List<String> listClassifiers()

Returns a list of all classifier models currently stored on the server.

Specified by:

listClassifiers in interface WekaService

Returns:

the list of classifier models

listClusterers

public List<String> listClusterers()

Returns a list of all clusterer models currently stored on the server.

Specified by:

listClusterers in interface WekaService

Returns:

the list of clusterer models

optimizeClassifierMultiSearch

public OptimizeReturnObject optimizeClassifierMultiSearch(String classifier,
                                                          List<String> searchParameters,
                                                          Dataset dataset,
                                                          String evaluation)

Optimizes the parameters of a base classifier using MultiSearch.

Specified by:

optimizeClassifierMultiSearch in interface WekaService

Parameters:

classifier - the base classifier

searchParameters - the search parameters for the optimization

dataset - the dataset to evaluate on

evaluation - the statistic to use for performance evaluation

Returns:

the result of the optimization

displayString

protected void displayString(Dataset dataset)

For outputting the dataset in debug mode.

Parameters:

dataset - the dataset to output