weka.filters.supervised.instance
Class SpreadSubsample

java.lang.Object
  weka.filters.Filter
      weka.filters.supervised.instance.SpreadSubsample

All Implemented Interfaces:

OptionHandler, java.io.Serializable, SupervisedFilter

public class SpreadSubsample

extends Filter

implements SupervisedFilter, OptionHandler

Produces a random subsample of a dataset. The original dataset must fit entirely in memory. This filter allows you to specify the maximum "spread" between the rarest and most common class. For example, you may specify that there be at most a 2:1 difference in class frequencies. When used in batch mode, subsequent batches are not resampled. Valid options are:

-S num
Specify the random number seed (default 1).

-M num
The maximum class distribution spread.
0 = no maximum spread, 1 = uniform distribution, 10 = allow at most a 10:1 ratio between the classes (default 0)

-X num
The maximum count for any class value.
(default 0 = unlimited)

-W
Adjust weights so that total weight per class is maintained. Individual instance weighting is not preserved. (default no weights adjustment)

Version:

$Revision: 1.3 $

Author:

Stuart Inglis (stuart@reeltwo.com)

See Also:

Serialized Form

Field Summary

private boolean	m_AdjustWeights True if instance weights will be adjusted to maintain total weight per class.
private double	m_DistributionSpread True if the first batch has been done
private boolean	m_FirstBatchDone True if the first batch has been done
private int	m_MaxCount The maximum count of any class
private int	m_RandomSeed The random number generator seed

Fields inherited from class weka.filters.Filter

m_NewBatch

Constructor Summary

SpreadSubsample()

Method Summary

java.lang.String	adjustWeightsTipText() Returns the tip text for this property
boolean	batchFinished() Signify that this batch of input to the filter is finished.
private void	createSubsample() Creates a subsample of the current set of input instances.
java.lang.String	distributionSpreadTipText() Returns the tip text for this property
boolean	getAdjustWeights() Returns true if instance weights will be adjusted to maintain total weight per class.
private int[]	getClassIndices() Creates an index containing the position where each class starts in the getInputFormat(). m_InputFormat must be sorted on the class attribute.
double	getDistributionSpread() Gets the value for the distribution spread
double	getMaxCount() Gets the value for the max count
java.lang.String[]	getOptions() Gets the current settings of the filter.
int	getRandomSeed() Gets the random number seed.
java.lang.String	globalInfo() Returns a string describing this filter
boolean	input(Instance instance) Input an instance for filtering.
java.util.Enumeration	listOptions() Returns an enumeration describing the available options.
static void	main(java.lang.String[] argv) Main method for testing this class.
java.lang.String	maxCountTipText() Returns the tip text for this property
java.lang.String	randomSeedTipText() Returns the tip text for this property
void	setAdjustWeights(boolean newAdjustWeights) Sets whether the instance weights will be adjusted to maintain total weight per class.
void	setDistributionSpread(double spread) Sets the value for the distribution spread
boolean	setInputFormat(Instances instanceInfo) Sets the format of the input instances.
void	setMaxCount(double maxcount) Sets the value for the max count
void	setOptions(java.lang.String[] options) Parses a list of options for this object.
void	setRandomSeed(int newSeed) Sets the random number seed.

Methods inherited from class weka.filters.Filter

batchFilterFile, bufferInput, copyStringValues, copyStringValues, filterFile, flushInput, getInputFormat, getInputStringIndex, getOutputFormat, getOutputStringIndex, getStringIndices, inputFormat, inputFormatPeek, isOutputFormatDefined, numPendingOutput, output, outputFormat, outputFormatPeek, outputPeek, push, resetQueue, setOutputFormat, useFilter

Methods inherited from class java.lang.Object

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

Field Detail

m_RandomSeed

private int m_RandomSeed

The random number generator seed

m_MaxCount

private int m_MaxCount

The maximum count of any class

m_FirstBatchDone

private boolean m_FirstBatchDone

True if the first batch has been done

m_DistributionSpread

private double m_DistributionSpread

True if the first batch has been done

m_AdjustWeights

private boolean m_AdjustWeights

True if instance weights will be adjusted to maintain total weight per class.

Constructor Detail

SpreadSubsample

public SpreadSubsample()

Method Detail

globalInfo

public java.lang.String globalInfo()

Returns a string describing this filter

Returns:

a description of the filter suitable for displaying in the explorer/experimenter gui

adjustWeightsTipText

public java.lang.String adjustWeightsTipText()

Returns the tip text for this property

Returns:

tip text for this property suitable for displaying in the explorer/experimenter gui

getAdjustWeights

public boolean getAdjustWeights()

Returns true if instance weights will be adjusted to maintain total weight per class.

Returns:

true if instance weights will be adjusted to maintain total weight per class.

setAdjustWeights

public void setAdjustWeights(boolean newAdjustWeights)

Sets whether the instance weights will be adjusted to maintain total weight per class.

Parameters:

newAdjustWeights -

listOptions

public java.util.Enumeration listOptions()

Returns an enumeration describing the available options.

Specified by:

listOptions in interface OptionHandler

Returns:

an enumeration of all the available options.

setOptions

public void setOptions(java.lang.String[] options)
                throws java.lang.Exception

Parses a list of options for this object. Valid options are:

-S num
Specify the random number seed (default 1).

-M num
The maximum class distribution spread.
0 = no maximum spread, 1 = uniform distribution, 10 = allow at most a 10:1 ratio between the classes (default 0)

-X num
The maximum count for any class value.
(default 0 = unlimited)

-W
Adjust weights so that total weight per class is maintained. Individual instance weighting is not preserved. (default no weights adjustment)

Specified by:

setOptions in interface OptionHandler

Parameters:

options - the list of options as an array of strings

Throws:

java.lang.Exception - if an option is not supported

getOptions

public java.lang.String[] getOptions()

Gets the current settings of the filter.

Specified by:

getOptions in interface OptionHandler

Returns:

an array of strings suitable for passing to setOptions

distributionSpreadTipText

public java.lang.String distributionSpreadTipText()

Returns the tip text for this property

Returns:

tip text for this property suitable for displaying in the explorer/experimenter gui

setDistributionSpread

public void setDistributionSpread(double spread)

Sets the value for the distribution spread

Parameters:

spread - the new distribution spread

getDistributionSpread

public double getDistributionSpread()

Gets the value for the distribution spread

Returns:

the distribution spread

maxCountTipText

public java.lang.String maxCountTipText()

Returns the tip text for this property

Returns:

tip text for this property suitable for displaying in the explorer/experimenter gui

setMaxCount

public void setMaxCount(double maxcount)

Sets the value for the max count

getMaxCount

public double getMaxCount()

Gets the value for the max count

Returns:

the max count

randomSeedTipText

public java.lang.String randomSeedTipText()

Returns the tip text for this property

Returns:

tip text for this property suitable for displaying in the explorer/experimenter gui

getRandomSeed

public int getRandomSeed()

Gets the random number seed.

Returns:

the random number seed.

setRandomSeed

public void setRandomSeed(int newSeed)

Sets the random number seed.

Parameters:

newSeed - the new random number seed.

setInputFormat

public boolean setInputFormat(Instances instanceInfo)
                       throws java.lang.Exception

Sets the format of the input instances.

Overrides:

setInputFormat in class Filter

Parameters:

instanceInfo - an Instances object containing the input instance structure (any instances contained in the object are ignored - only the structure is required).

Returns:

true if the outputFormat may be collected immediately

Throws:

UnassignedClassException - if no class attribute has been set.

UnsupportedClassTypeException - if the class attribute is not nominal.

java.lang.Exception - if the inputFormat can't be set successfully

input

public boolean input(Instance instance)

Input an instance for filtering. Filter requires all training instances be read before producing output.

Overrides:

input in class Filter

Parameters:

instance - the input instance

Returns:

true if the filtered instance may now be collected with output().

Throws:

java.lang.IllegalStateException - if no input structure has been defined

batchFinished

public boolean batchFinished()

Signify that this batch of input to the filter is finished. If the filter requires all instances prior to filtering, output() may now be called to retrieve the filtered instances.

Overrides:

batchFinished in class Filter

Returns:

true if there are instances pending output

Throws:

java.lang.IllegalStateException - if no input structure has been defined

createSubsample

private void createSubsample()

Creates a subsample of the current set of input instances. The output instances are pushed onto the output queue for collection.

getClassIndices

private int[] getClassIndices()

Creates an index containing the position where each class starts in the getInputFormat(). m_InputFormat must be sorted on the class attribute.

main

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

Main method for testing this class.

Parameters:

argv - should contain arguments to the filter: use -h for help