marf.Storage
Class TrainingSample

java.lang.Object
  marf.Storage.TrainingSample

All Implemented Interfaces:

java.io.Serializable, java.lang.Cloneable, ITrainingSample

Direct Known Subclasses:

Cluster

public class TrainingSample
extends java.lang.Object
implements ITrainingSample

TrainingSample contains one item in the training set. Each training sample consists of the feature vector plus information describing that feature vector. Has been extracted from TrainingSet in 0.3.0. TODO: fix CSV dumps.

$Id: TrainingSample.java,v 1.15 2008/02/20 07:29:06 mokhov Exp $

Since:

0.0.1

Version:

$Revision: 1.15 $

Author:

Stephen Sinclair, Serguei Mokhov

See Also:

Serialized Form

Field Summary

protected double[]	adDataVector Array representing either a feature vector, mean vector describing the cluster, or a median vector.
protected int	iSubjectID Which subject this feature vector is associated with.
protected java.util.Vector	oFilenames A list of filenames that were used in training for this sample.
private static long	serialVersionUID For serialization versioning.

Constructor Summary

TrainingSample()
Default training sample constructor.

TrainingSample(TrainingSample poTrainingSample)
Copy-constructor.

Method Summary

boolean	addFeatureVector(double[] padFeatureVector, java.lang.String pstrFilename, int piSubjectID) Fully equivalent to setFeatureVector().
boolean	addFilename(java.lang.String pstrFilename) Adds a filename to the training sample.
java.lang.Object	clone() Implements Cloneable interface for the TrainingSample object.
void	dumpCSV(java.io.BufferedWriter poBufferedWriter) Writes one training sample to a CSV file.
boolean	existsFilename(java.lang.String pstrFilename) Checks existence of the file in the training sample.
double[]	getDataVector() Retrieves the data vector.
static java.lang.String	getMARFSourceCodeRevision() Returns source code revision information.
int	getMeanCount() Retrieves current mean count.
double[]	getMeanVector() Simply retrieves the data vector.
double[]	getMedianVector() Simply retrieves the data vector.
int	getSubjectID() Retrieves Subject ID of a particular training sample.
void	restoreCSV(java.io.BufferedReader poBufferedReader) Retrieve one training sample from a CSV file.
void	setDataVector(double[] padDataVector) Sets new mean vector.
boolean	setFeatureVector(double[] padFeatureVector, java.lang.String pstrFilename, int piSubjectID)
void	setFilename(java.lang.String pstrFilename) Sets a filename of the training sample.
void	setSubjectID(int piSubjectID) Sets new Subject ID.
int	size()
java.lang.String	toString() Provides string representation of the training sample data.

Methods inherited from class java.lang.Object

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

Field Detail

iSubjectID

protected int iSubjectID

Which subject this feature vector is associated with. This is an application-independent ID within MARF to distinguish from other subjects. Typical subjects may include speakers, languages instruments, emotions, etc. that a give application is taken care of.

adDataVector

protected double[] adDataVector

Array representing either a feature vector, mean vector describing the cluster, or a median vector.

oFilenames

protected java.util.Vector oFilenames

A list of filenames that were used in training for this sample. Used to avoid duplicate training on the same filename.

serialVersionUID

private static final long serialVersionUID

For serialization versioning. When adding new members or make other structural changes regenerate this number with the serialver tool that comes with JDK.

Since:

0.3.0.4

See Also:

Constant Field Values

Constructor Detail

TrainingSample

public TrainingSample()

Default training sample constructor. Explicitly appeared in 0.3.0.5.

Since:

0.3.0.5

TrainingSample

public TrainingSample(TrainingSample poTrainingSample)

Copy-constructor.

Parameters:

poTrainingSample - TrainingSample object to copy

Since:

0.3.0.5

Method Detail

setFeatureVector

public boolean setFeatureVector(double[] padFeatureVector,
                                java.lang.String pstrFilename,
                                int piSubjectID)

Specified by:

setFeatureVector in interface ITrainingSample

Returns:

addFeatureVector

public boolean addFeatureVector(double[] padFeatureVector,
                                java.lang.String pstrFilename,
                                int piSubjectID)

Description copied from interface: ITrainingSample

Fully equivalent to setFeatureVector().

Specified by:

addFeatureVector in interface ITrainingSample

Returns:

See Also:

ITrainingSample.setFeatureVector(double[], String, int)

setFilename

public void setFilename(java.lang.String pstrFilename)

Description copied from interface: ITrainingSample

Sets a filename of the training sample. Always set the first element of the list of filenames.

Specified by:

setFilename in interface ITrainingSample

Parameters:

pstrFilename - filename to set

addFilename

public boolean addFilename(java.lang.String pstrFilename)

Adds a filename to the training sample. The method is protected and can be made public by the extending classes that allow more than one filename.

Specified by:

addFilename in interface ITrainingSample

Parameters:

pstrFilename - filename to add

Returns:

false if the filename is already there; true otherwise

Since:

0.3.0.6

See Also:

existsFilename(String)

existsFilename

public boolean existsFilename(java.lang.String pstrFilename)

Description copied from interface: ITrainingSample

Checks existence of the file in the training sample. Serves as an indication that we already trained on the given file.

Specified by:

existsFilename in interface ITrainingSample

Parameters:

pstrFilename - filename to check

Returns:

true if the filename is there; false if not

getSubjectID

public final int getSubjectID()

Description copied from interface: ITrainingSample

Retrieves Subject ID of a particular training sample.

Specified by:

getSubjectID in interface ITrainingSample

Returns:

int ID

getDataVector

public final double[] getDataVector()

Description copied from interface: ITrainingSample

Retrieves the data vector.

Specified by:

getDataVector in interface ITrainingSample

Returns:

array of doubles representing the data for this subject

setSubjectID

public void setSubjectID(int piSubjectID)

Description copied from interface: ITrainingSample

Sets new Subject ID.

Specified by:

setSubjectID in interface ITrainingSample

Parameters:

piSubjectID - integer ID

setDataVector

public void setDataVector(double[] padDataVector)

Description copied from interface: ITrainingSample

Sets new mean vector.

Specified by:

setDataVector in interface ITrainingSample

Parameters:

padDataVector - double array representing the mean vector

dumpCSV

public void dumpCSV(java.io.BufferedWriter poBufferedWriter)
             throws StorageException

Writes one training sample to a CSV file.

Parameters:

poBufferedWriter - BufferedWriter

Throws:

StorageException - in case of I/O or otherwise error

restoreCSV

public void restoreCSV(java.io.BufferedReader poBufferedReader)
                throws StorageException

Retrieve one training sample from a CSV file. TODO: FIX, for it's BROKEN.

Parameters:

poBufferedReader - BufferedReader

Throws:

StorageException - in case of I/O or otherwise error

getMeanCount

public int getMeanCount()

Description copied from interface: ITrainingSample

Retrieves current mean count.

Specified by:

getMeanCount in interface ITrainingSample

Returns:

mean count

Since:

0.3.0.6

See Also:

ITrainingSample.getMeanCount()

getMeanVector

public double[] getMeanVector()

Simply retrieves the data vector. Internally calls getDataVector().

Specified by:

getMeanVector in interface ITrainingSample

Returns:

array of doubles representing the mean for that cluster

Since:

0.3.0.6

See Also:

getDataVector()

getMedianVector

public double[] getMedianVector()

Simply retrieves the data vector. Internally calls getDataVector().

Specified by:

getMedianVector in interface ITrainingSample

Returns:

array of doubles representing the mean for that cluster

Since:

0.3.0.6

See Also:

getDataVector()

size

public int size()

Specified by:

size in interface ITrainingSample

Returns:

Since:

0.3.0.6

See Also:

ITrainingSample.size()

clone

public java.lang.Object clone()

Implements Cloneable interface for the TrainingSample object.

Overrides:

clone in class java.lang.Object

Since:

0.3.0.5

See Also:

Object.clone()

toString

public java.lang.String toString()

Provides string representation of the training sample data.

Overrides:

toString in class java.lang.Object

Since:

0.3.0.6

See Also:

Object.toString()

getMARFSourceCodeRevision

public static java.lang.String getMARFSourceCodeRevision()

Returns source code revision information.

Returns:

revision string

Since:

0.3.0.2