weka.experiment

Class RemoteExperiment

java.lang.Object

weka.experiment.Experiment

weka.experiment.RemoteExperiment

All Implemented Interfaces:

java.io.Serializable, OptionHandler, RevisionHandler

public class RemoteExperiment
extends Experiment

Holds all the necessary configuration information for a distributed experiment. This object is able to be serialized for storage on disk.

This class is experimental at present. Has been tested using CSVResultListener (sending results to standard out) and DatabaseResultListener (InstantDB + RmiJdbc bridge).

Getting started:

Start InstantDB (with the RMI bridge) on some machine. If using java2 then specify -Djava.security.policy=db.policy to the virtual machine. Where db.policy is as follows:

 grant {
   permission java.security.AllPermission;
 };
 

Start RemoteEngine servers on x machines as per the instructons in the README_Experiment_Gui file. There must be a DatabaseUtils.props in either the HOME or current directory of each machine, listing all necessary jdbc drivers.

The machine where a RemoteExperiment is started must also have a copy of DatabaseUtils.props listing the URL to the machine where the database server is running (RmiJdbc + InstantDB).

Here is an example of starting a RemoteExperiment:

 
 java -Djava.rmi.server.codebase=file:/path to weka classes/ \
 weka.experiment.RemoteExperiment -L 1 -U 10 \
 -T /home/ml/datasets/UCI/iris.arff \
 -D "weka.experiment.DatabaseResultListener" \
 -P "weka.experiment.RandomSplitResultProducer" \
 -h rosebud.cs.waikato.ac.nz -h blackbird.cs.waikato.ac.nz -r -- \
 -W weka.experiment.ClassifierSplitEvaluator -- \
 -W weka.classifiers.bayes.NaiveBayes
 
 

The "codebase" property tells rmi where to serve up weka classes from. This can either be a file url (as long as a shared file system is being used that is accessable by the remoteEngine servers), or http url (which of course supposes that a web server is running and you have put your weka classes somewhere that is web accessable). If using a file url the trailing "/" is *most* important unless the weka classes are in a jar file.

Valid options are:

 -L <num>
  The lower run number to start the experiment from.
  (default 1)
 

 -U <num>
  The upper run number to end the experiment at (inclusive).
  (default 10)
 

 -T <arff file>
  The dataset to run the experiment on.
  (required, may be specified multiple times)
 

 -P <class name>
  The full class name of a ResultProducer (required).
  eg: weka.experiment.RandomSplitResultProducer
 

 -D <class name>
  The full class name of a ResultListener (required).
  eg: weka.experiment.CSVResultListener
 

 -N <string>
  A string containing any notes about the experiment.
  (default none)
 

 Options specific to result producer weka.experiment.RandomSplitResultProducer:
 

 -P <percent>
  The percentage of instances to use for training.
  (default 66)
 

 -D
 Save raw split evaluator output.
 

 -O <file/directory name/path>
  The filename where raw output will be stored.
  If a directory name is specified then then individual
  outputs will be gzipped, otherwise all output will be
  zipped to the named file. Use in conjuction with -D. (default splitEvalutorOut.zip)
 

 -W <class name>
  The full class name of a SplitEvaluator.
  eg: weka.experiment.ClassifierSplitEvaluator
 

 -R
  Set when data is not to be randomized and the data sets' size.
  Is not to be determined via probabilistic rounding.
 

 Options specific to split evaluator weka.experiment.ClassifierSplitEvaluator:
 

 -W <class name>
  The full class name of the classifier.
  eg: weka.classifiers.bayes.NaiveBayes
 

 -C <index>
  The index of the class for which IR statistics
  are to be output. (default 1)
 

 -I <index>
  The index of an attribute to output in the
  results. This attribute should identify an
  instance in order to know which instances are
  in the test set of a cross validation. if 0
  no output (default 0).
 

 -P
  Add target and prediction columns to the result
  for each fold.
 

 Options specific to classifier weka.classifiers.rules.ZeroR:
 

 -D
  If set, classifier is run in debug mode and
  may output additional info to the console
 

Version:

$Revision: 12590 $

Author:

Mark Hall (mhall@cs.waikato.ac.nz)

See Also:

Serialized Form

Field Summary

Fields inherited from class weka.experiment.Experiment

FILE_EXTENSION

Constructor Summary

Constructors

Constructor and Description
RemoteExperiment() Construct a new RemoteExperiment using an empty Experiment as base Experiment
RemoteExperiment(Experiment base) Construct a new RemoteExperiment using a base Experiment

Method Summary

Modifier and Type	Method and Description
void	abortExperiment() Set the abort flag
void	addRemoteExperimentListener(RemoteExperimentListener r) Add an object to the list of those interested in recieving update information from the RemoteExperiment
void	addRemoteHost(java.lang.String hostname) Add a host name to the list of remote hosts
void	advanceCounters() overides the one in Experiment
Experiment	getBaseExperiment() Get the base experiment used by this remote experiment
javax.swing.DefaultListModel	getRemoteHosts() Get the list of remote host names
java.lang.String	getRevision() Returns the revision string.
boolean	getSplitByDataSet() Returns true if sub experiments are to be created on the basis of data set.
boolean	getSplitByProperty() Returns true if sub experiments are to be created on the basis of property.
void	initialize() Prepares a remote experiment for running, creates sub experiments
void	launchNext(int wexp, int ah) Launch a sub experiment on a remote host
static void	main(java.lang.String[] args) Configures/Runs the Experiment from the command line.
void	nextIteration() Overides the one in Experiment
void	postProcess() overides the one in Experiment
void	runExperiment() Overides runExperiment in Experiment
void	setBaseExperiment(Experiment base) Set the base experiment.
void	setDatasets(javax.swing.DefaultListModel ds) Set the datasets to use in the experiment
void	setNotes(java.lang.String newNotes) Set the user notes.
void	setPropertyArray(java.lang.Object newPropArray) Sets the array of values to set the custom property to.
void	setPropertyPath(PropertyNode[] newPropertyPath) Sets the path of properties taken to get to the custom property to iterate over.
void	setRemoteHosts(javax.swing.DefaultListModel list) Set the list of remote host names
void	setResultListener(ResultListener newResultListener) Sets the result listener where results will be sent.
void	setResultProducer(ResultProducer newResultProducer) Set the result producer used for the current experiment.
void	setRunLower(int newRunLower) Set the lower run number for the experiment.
void	setRunUpper(int newRunUpper) Set the upper run number for the experiment.
void	setSplitByDataSet(boolean sd) Set whether sub experiments are to be created on the basis of data set.
void	setSplitByProperty(boolean sd) Set whether sub experiments are to be created on the basis of property.
void	setUsePropertyIterator(boolean newUsePropertyIterator) Sets whether the custom property iterator should be used.
java.lang.String	toString() Overides toString in Experiment

Methods inherited from class weka.experiment.Experiment

classFirst, getAdvanceDataSetFirst, getCurrentDatasetNumber, getCurrentPropertyNumber, getCurrentRunNumber, getDatasets, getNotes, getOptions, getPropertyArray, getPropertyArrayLength, getPropertyArrayValue, getPropertyPath, getResultListener, getResultProducer, getRunLower, getRunUpper, getUsePropertyIterator, hasMoreIterations, listOptions, read, runExperiment, setAdvanceDataSetFirst, setOptions, write

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Methods inherited from interface weka.core.OptionHandler

makeCopy

Constructor Detail

RemoteExperiment

public RemoteExperiment()
                 throws java.lang.Exception

Construct a new RemoteExperiment using an empty Experiment as base Experiment

Throws:

java.lang.Exception - if the base experiment is null

RemoteExperiment

public RemoteExperiment(Experiment base)
                 throws java.lang.Exception

Construct a new RemoteExperiment using a base Experiment

Parameters:

base - the base experiment to use

Throws:

java.lang.Exception - if the base experiment is null

Method Detail

getSplitByDataSet

public boolean getSplitByDataSet()

Returns true if sub experiments are to be created on the basis of data set.

Returns:

a boolean value indicating whether sub experiments are to be created on the basis of data set (true) or run number (false).

setSplitByDataSet

public void setSplitByDataSet(boolean sd)

Set whether sub experiments are to be created on the basis of data set.

Parameters:

sd - true if sub experiments are to be created on the basis of data set. Otherwise sub experiments are created on the basis of run number.

getSplitByProperty

public boolean getSplitByProperty()

Returns true if sub experiments are to be created on the basis of property.

Returns:

a boolean value indicating whether sub experiments are to be created on the basis of data set (true) or run number (false).

setSplitByProperty

public void setSplitByProperty(boolean sd)

Set whether sub experiments are to be created on the basis of property.

Parameters:

sd - true if sub experiments are to be created on the basis of data set. Otherwise sub experiments are created on the basis of run number.

addRemoteExperimentListener

public void addRemoteExperimentListener(RemoteExperimentListener r)

Add an object to the list of those interested in recieving update information from the RemoteExperiment

Parameters:

r - a listener

getBaseExperiment

public Experiment getBaseExperiment()

Get the base experiment used by this remote experiment

Returns:

the base experiment

setBaseExperiment

public void setBaseExperiment(Experiment base)
                       throws java.lang.Exception

Set the base experiment. A sub experiment will be created for each run in the base experiment.

Parameters:

base - the base experiment to use.

Throws:

java.lang.Exception - if supplied base experiment is null

setNotes

public void setNotes(java.lang.String newNotes)

Set the user notes.

Overrides:

setNotes in class Experiment

Parameters:

newNotes - New user notes.

setRunLower

public void setRunLower(int newRunLower)

Set the lower run number for the experiment.

Overrides:

setRunLower in class Experiment

Parameters:

newRunLower - the lower run number for the experiment.

setRunUpper

public void setRunUpper(int newRunUpper)

Set the upper run number for the experiment.

Overrides:

setRunUpper in class Experiment

Parameters:

newRunUpper - the upper run number for the experiment.

setResultListener

public void setResultListener(ResultListener newResultListener)

Sets the result listener where results will be sent.

Overrides:

setResultListener in class Experiment

Parameters:

newResultListener - the result listener where results will be sent.

setResultProducer

public void setResultProducer(ResultProducer newResultProducer)

Set the result producer used for the current experiment.

Overrides:

setResultProducer in class Experiment

Parameters:

newResultProducer - result producer to use for the current experiment.

setDatasets

public void setDatasets(javax.swing.DefaultListModel ds)

Set the datasets to use in the experiment

Overrides:

setDatasets in class Experiment

Parameters:

ds - the list of datasets to use

setUsePropertyIterator

public void setUsePropertyIterator(boolean newUsePropertyIterator)

Sets whether the custom property iterator should be used.

Overrides:

setUsePropertyIterator in class Experiment

Parameters:

newUsePropertyIterator - true if so

setPropertyPath

public void setPropertyPath(PropertyNode[] newPropertyPath)

Sets the path of properties taken to get to the custom property to iterate over.

Overrides:

setPropertyPath in class Experiment

Parameters:

newPropertyPath - an array of PropertyNodes

setPropertyArray

public void setPropertyArray(java.lang.Object newPropArray)

Sets the array of values to set the custom property to.

Overrides:

setPropertyArray in class Experiment

Parameters:

newPropArray - a value of type Object which should be an array of the appropriate values.

initialize

public void initialize()
                throws java.lang.Exception

Prepares a remote experiment for running, creates sub experiments

Overrides:

initialize in class Experiment

Throws:

java.lang.Exception - if an error occurs

abortExperiment

public void abortExperiment()

Set the abort flag

launchNext

public void launchNext(int wexp,
                       int ah)

Launch a sub experiment on a remote host

Parameters:

wexp - the index of the sub experiment to launch

ah - the index of the available host to launch on

nextIteration

public void nextIteration()
                   throws java.lang.Exception

Overides the one in Experiment

Overrides:

nextIteration in class Experiment

Throws:

java.lang.Exception - never throws an exception

advanceCounters

public void advanceCounters()

overides the one in Experiment

Overrides:

advanceCounters in class Experiment

postProcess

public void postProcess()

overides the one in Experiment

Overrides:

postProcess in class Experiment

addRemoteHost

public void addRemoteHost(java.lang.String hostname)

Add a host name to the list of remote hosts

Parameters:

hostname - the host name to add to the list

getRemoteHosts

public javax.swing.DefaultListModel getRemoteHosts()

Get the list of remote host names

Returns:

the list of remote host names

setRemoteHosts

public void setRemoteHosts(javax.swing.DefaultListModel list)

Set the list of remote host names

Parameters:

list - the list of remote host names

toString

public java.lang.String toString()

Overides toString in Experiment

Overrides:

toString in class Experiment

Returns:

a description of this remote experiment

runExperiment

public void runExperiment()

Overides runExperiment in Experiment

Overrides:

runExperiment in class Experiment

getRevision

public java.lang.String getRevision()

Returns the revision string.

Specified by:

getRevision in interface RevisionHandler

Overrides:

getRevision in class Experiment

Returns:

the revision

main

public static void main(java.lang.String[] args)

Configures/Runs the Experiment from the command line.

Parameters:

args - command line arguments to the Experiment.