magpie.data.utilities.splitters

Class NonZeroClassificationSplitter

java.lang.Object

magpie.data.utilities.splitters.BaseDatasetSplitter

magpie.data.utilities.splitters.NonZeroClassificationSplitter

All Implemented Interfaces:

java.io.Serializable, java.lang.Cloneable, Commandable, Options, Printable

public class NonZeroClassificationSplitter
extends BaseDatasetSplitter

Split data based on whether an entry is likely to have a non-zero class variable. If the probability of the entry having a class variable equal to zero is past a certain threshold, it is assigned to subset #0.

Usage: <threshold> $<model>
threshold: Probability of having a class == 0 on which to split
model: AbstractClassifier to use for partitioning data. Must distinguish between two classes

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected BaseModel	Clfr Model to use for splitting
protected double	ProbabilityThreshold Probability of being of class #0 below which entry gets placed in entry #1

Constructor Summary

Constructors

Constructor and Description
NonZeroClassificationSplitter() Create a new instances that uses ZeroR to classify.

Method Summary

Modifier and Type	Method and Description
NonZeroClassificationSplitter	clone()
java.util.List<java.lang.String>	getSplitNames() Get the names of the splits this splitter creates
protected java.util.List<java.lang.String>	getSplitterDetails(boolean htmlFormat) Get details of splitter.
int[]	label(Dataset D) Given a dataset, determine which subset each entry should belong to.
java.lang.String	printUsage() Print out required format for options.
void	setModel(BaseModel model) Define the model used to split data.
void	setOptions(java.util.List Options) Set any options for this object.
void	setProbabilityThreshold(double ProbabilityThreshold) Define the probability of being in class 0 on which to partition data.
void	train(Dataset TrainingSet) Train a classifier for determining which entries have a non-zero class

Methods inherited from class magpie.data.utilities.splitters.BaseDatasetSplitter

about, printCommand, printDescription, runCommand, split, split

Methods inherited from class java.lang.Object

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

Field Detail

Clfr

protected BaseModel Clfr

Model to use for splitting

ProbabilityThreshold

protected double ProbabilityThreshold

Probability of being of class #0 below which entry gets placed in entry #1

Constructor Detail

NonZeroClassificationSplitter

public NonZeroClassificationSplitter()
                              throws java.lang.Exception

Create a new instances that uses ZeroR to classify.

Throws:

java.lang.Exception

Method Detail

clone

public NonZeroClassificationSplitter clone()

Overrides:

clone in class BaseDatasetSplitter

setOptions

public void setOptions(java.util.List Options)
                throws java.lang.Exception

Description copied from interface: Options

Set any options for this object. If parsing the options fails, return an Exception with the usage as its message.

Parameters:

Options - Array of options as Objects - can be null

Throws:

java.lang.Exception - if problem with inputs

printUsage

public java.lang.String printUsage()

Description copied from interface: Options

Print out required format for options. For consistency, should start with "Usage: ".

Returns:

Usage for this function

setProbabilityThreshold

public void setProbabilityThreshold(double ProbabilityThreshold)

Define the probability of being in class 0 on which to partition data.

Parameters:

ProbabilityThreshold - Desired threshold

setModel

public void setModel(BaseModel model)
              throws java.lang.Exception

Define the model used to split data. Should implement AbstractClassifier

Parameters:

model - Desired model (will be cloned)

Throws:

java.lang.Exception

label

public int[] label(Dataset D)

Description copied from class: BaseDatasetSplitter

Given a dataset, determine which subset each entry should belong to. Please ensure that D remains unaltered (clone if you must)

Specified by:

label in class BaseDatasetSplitter

Parameters:

D - Dataset to be labeled

Returns:

A list of nonnegative integers marking which subset to split each entry into

train

public void train(Dataset TrainingSet)

Train a classifier for determining which entries have a non-zero class

Specified by:

train in class BaseDatasetSplitter

Parameters:

TrainingSet -

getSplitterDetails

protected java.util.List<java.lang.String> getSplitterDetails(boolean htmlFormat)

Description copied from class: BaseDatasetSplitter

Get details of splitter. Returns a list of options that can be set by the user. Used with BaseDatasetSplitter.printDescription(boolean).

Specified by:

getSplitterDetails in class BaseDatasetSplitter

Parameters:

htmlFormat - Whether to format results with HTML

Returns:

List of options describing the model

getSplitNames

public java.util.List<java.lang.String> getSplitNames()

Description copied from class: BaseDatasetSplitter

Get the names of the splits this splitter creates

Specified by:

getSplitNames in class BaseDatasetSplitter

Returns:

Names of each group