uk.ac.ed.inf.enzml.mulan
Class MulanArff

java.lang.Object
  uk.ac.ed.inf.utils.Initialised
      uk.ac.ed.inf.enzml.weka.Arff
          uk.ac.ed.inf.enzml.mulan.MulanArff

All Implemented Interfaces:

uk.ac.ed.inf.utils.Initialisable

Direct Known Subclasses:

AttributesFilteredArff

public class MulanArff
extends Arff

Generates an Attribute Relation Format File (ARFF) in Mulan format (multiple class labels per instance). The class attributes must be in a separate Mulan XML file (either flat or hierarchical) to distinguish them from the normal attributes. Call structure: the MulanArffGenerator creates a DataSetManager which checks the input (DataSetChecker), loads the data (MulanDataSetDbLoader), creates the instances (MulanDataSetGenerator) and writes them as ARFF file (DataSetWriter).

Version:

10 May 2010

Author:

Luna De Ferrari luna.deferrari-at-ed.ac.uk

Field Summary

boolean

m_memoryAbundant if plenty of memory available, also load each instance in weka instances, otherwise only write them to file one by one

Fields inherited from class uk.ac.ed.inf.utils.Initialised

m_initialisationIsCorrect

Constructor Summary

MulanArff(java.lang.String dataPropsFileName, java.lang.String resultsPropsFileName)

MulanArff(java.lang.String dataPropsFileName, java.lang.String resultsPropsFileName, boolean memoryAbundant)

Method Summary

MulanDataSetManager	getDataSetManager()
weka.core.Instances	getInstances()
java.util.Vector<java.lang.String>	getInstancesNames()
java.lang.String	getXmlPath()
static java.lang.String	getXMLpathFromArffPath(java.lang.String arffPath) Generates the xml file path: same name+path as .arff file, but .xml termination
static void	main(java.lang.String[] args)
static int[]	run(java.lang.String[] arffPaths, java.lang.String machineLearningProps) Generates a number of ARFF files (the database the records are written is the same for all ARFF files)
static int	run(java.lang.String arffPropsPath, java.lang.String machineLearningProps) Generates the ARFF (and XML) files and returns the identifier of the arff record in the database
void	setDataSetManager()

Methods inherited from class uk.ac.ed.inf.enzml.weka.Arff

checkAndInitialise, fillArffButDontWrite, generateAndWriteArffFile, generateArff, getArffPath, getArffRecordId, getDataArffProperties, getDataDbManager, getDataPropsFilePath, getDbConn, getDbReader, getResultsDbManager, memoryIsAbundant, setArffRecordId

Methods inherited from class uk.ac.ed.inf.utils.Initialised

initialisationIsCorrect, setInitialisationIsCorrect

Methods inherited from class java.lang.Object

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

Field Detail

m_memoryAbundant

public boolean m_memoryAbundant

if plenty of memory available, also load each instance in weka instances, otherwise only write them to file one by one

Constructor Detail

MulanArff

public MulanArff(java.lang.String dataPropsFileName,
                 java.lang.String resultsPropsFileName)

MulanArff

public MulanArff(java.lang.String dataPropsFileName,
                 java.lang.String resultsPropsFileName,
                 boolean memoryAbundant)

Method Detail

getDataSetManager

public MulanDataSetManager getDataSetManager()

Overrides:

getDataSetManager in class Arff

getInstancesNames

public java.util.Vector<java.lang.String> getInstancesNames()

getXmlPath

public java.lang.String getXmlPath()

getInstances

public weka.core.Instances getInstances()

Overrides:

getInstances in class Arff

setDataSetManager

public void setDataSetManager()

Overrides:

setDataSetManager in class Arff

getXMLpathFromArffPath

public static java.lang.String getXMLpathFromArffPath(java.lang.String arffPath)

Generates the xml file path: same name+path as .arff file, but .xml termination

Parameters:

arffPath - the path to the arff file

Returns:

the same filename+directorypath of the arff file but with the termination .xml instead of .arff

main

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

run

public static int run(java.lang.String arffPropsPath,
                      java.lang.String machineLearningProps)

Generates the ARFF (and XML) files and returns the identifier of the arff record in the database

Parameters:

arffPropsPath - the path of the ARFF properties file (queries that define the data to be included)

machineLearningProps - the database connection to the database containing the arff file table, to save the arff file details.

Returns:

the identifier of the generated ARFF file in the database table

run

public static int[] run(java.lang.String[] arffPaths,
                        java.lang.String machineLearningProps)

Generates a number of ARFF files (the database the records are written is the same for all ARFF files)

Parameters:

arffPaths - a list of arff properties files (containing the data queries)

machineLearningProps - the database connection properties file to write the arff records too

Returns:

a list of arff records identifiers in the database arff table