modnlp.tc.dstruct

Class TCProbabilityModel

java.lang.Object

modnlp.tc.dstruct.TCProbabilityModel

All Implemented Interfaces:

java.io.Serializable

Direct Known Subclasses:

BVProbabilityModel

public abstract class TCProbabilityModel
extends java.lang.Object
implements java.io.Serializable

Define methodos to be implemented by all probability models for text categorisation. This is meant as an abstraction over probability models in which events are described as sets of documents (e.g. BVProbabilityModel, which implements the model described in [Sebastiani, 02] and elsewhere), and models where events are sets of token placeholders (as in [Mitchell, 97], for instance). It is assumed that TCProbabilityModel will be serialised and seved to disk for later use (e.g. by a classifer), hence the extension of Serializable and the fields and methods to store and retrieve creation information. TCProbabilityModels will be used typically in conjunction with TCInvertedIndex objects.

See Also:

TCInvertedIndex, Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
TCInvertedIndex	invertedIndex All TC probability models must be based on an inverted index
static byte	LAPLACE
static byte	NOSMOOTHING

Constructor Summary

Constructors

Constructor and Description
TCProbabilityModel()

Method Summary

Methods

Modifier and Type	Method and Description
java.lang.String	getCreationInfo() Get the value of creationInfo.
java.lang.String	getCreator() Get the value of creationInfo.
java.lang.String[]	getCreatorArgs() Get the value of creatorArgs.
java.lang.String	getCreatorArgsCSV() Get the value of creatorArgs as list of comma-separated values.
abstract Probabilities	getProbabilities(java.lang.String term, java.lang.String category) Get a summary of probabilities associated with term and category
byte	getSmoothingType()
void	setCreator(java.lang.String v) Set the value of creationInfo.
void	setCreatorArgs(java.lang.String[] v) Set the value of creatorArgs.
void	setSmoothingType(byte s)

Methods inherited from class java.lang.Object

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

Field Detail

NOSMOOTHING

public static final byte NOSMOOTHING

See Also:

Constant Field Values

LAPLACE

public static final byte LAPLACE

See Also:

Constant Field Values

invertedIndex

public TCInvertedIndex invertedIndex

All TC probability models must be based on an inverted index

Constructor Detail

TCProbabilityModel

public TCProbabilityModel()

Method Detail

getProbabilities

public abstract Probabilities getProbabilities(java.lang.String term,
                             java.lang.String category)

Get a summary of probabilities associated with term and category

Parameters:

term - a String value

category - a String value

Returns:

a Probabilities value

getCreator

public java.lang.String getCreator()

Get the value of creationInfo.

Returns:

value of creationInfo.

setCreator

public void setCreator(java.lang.String v)

Set the value of creationInfo.

Parameters:

v - Value to assign to creationInfo.

getCreatorArgs

public java.lang.String[] getCreatorArgs()

Get the value of creatorArgs.

Returns:

value of creatorArgs.

getCreatorArgsCSV

public java.lang.String getCreatorArgsCSV()

Get the value of creatorArgs as list of comma-separated values.

Returns:

value of creatorArgs.

setCreatorArgs

public void setCreatorArgs(java.lang.String[] v)

Set the value of creatorArgs.

Parameters:

v - Value to assign to creatorArgs.

getCreationInfo

public java.lang.String getCreationInfo()

Get the value of creationInfo.

Returns:

value of creationInfo.

getSmoothingType

public byte getSmoothingType()

setSmoothingType

public void setSmoothingType(byte s)