org.terrier.matching.models
Class DFRWeightingModel

java.lang.Object
  org.terrier.matching.models.WeightingModel
      org.terrier.matching.models.DFRWeightingModel

All Implemented Interfaces:

java.io.Serializable, java.lang.Cloneable, Model

public class DFRWeightingModel
extends WeightingModel

This class implements a modular Divergence from Randomness weighting model. Components in the model are specified individually, with many implementations provided: org.terrier.matching.models.basicmodel; org.terrier.matching.models.aftereffect; and org.terrier.matching.models.normalisation. A class name for each component should be specific in parenthesis after DFRWeightingModel. Moreover, besides the basic model, the other two components can be either specified or disabled.

Examples:

• DFRWeightingModel(P,L,2) is equivalent to PL2
• DFRWeightingModel(P,L) removes the term frequency normalisastion component from PL2

Author:

Ben He

See Also:

Serialized Form

Field Summary

protected AfterEffect	afterEffect The applied model for after effect (aka.
protected java.lang.String	AFTEREFFECT_PREFIX The prefix of the package of the first normalisation methods by after effect.
protected BasicModel	basicModel The applied basic model for randomness.
protected java.lang.String	BASICMODEL_PREFIX The prefix of the package of the basic models for randomness.
protected boolean	ENABLE_AFTEREFFECT A boolean that indicates if the first normalisation by after effect is enabled.
protected boolean	ENABLE_NORMALISATION A boolean that indicates if the frequency normalisation is enabled.
protected static org.apache.log4j.Logger	logger
protected Normalisation	normalisation The applied frequency normalisation method.
protected java.lang.String	NORMALISATION_PREFIX The prefix of the package of the frequency normalisation methods.
protected double	parameter The parameter of the frequency normalisation component.

Fields inherited from class org.terrier.matching.models.WeightingModel

averageDocumentLength, c, documentFrequency, i, keyFrequency, numberOfDocuments, numberOfPointers, numberOfTokens, numberOfUniqueTerms, termFrequency

Constructor Summary

DFRWeightingModel(java.lang.String[] components)
The default constructor.

Method Summary

java.lang.String	getInfo() Returns the name of the model.
double	getParameter() Return the parameter set by setParameter()
protected void	initialise(java.lang.String basicModelName, java.lang.String afterEffectName, java.lang.String normalisationName) Initialise the components in the DFR model.
protected void	initialise(java.lang.String basicModelName, java.lang.String afterEffectName, java.lang.String normalisationName, double _parameter) Initialise the components in the DFR model.
double	score(double tf, double docLength) Compute a weight for a term in a document.
double	score(double tf, double docLength, double documentFrequency, double termFrequency, double queryTermWeight) Compute a weight for a term in a document.
void	setCollectionStatistics(CollectionStatistics _cs) Sets collection statistics
void	setEntryStatistics(EntryStatistics _es) Sets entry statistics.
void	setParameter(double value) Set the frequency normalisation parameter.

Methods inherited from class org.terrier.matching.models.WeightingModel

clone, getOverflowed, prepare, score, setAverageDocumentLength, setDocumentFrequency, setKeyFrequency, setNumberOfDocuments, setNumberOfPointers, setNumberOfTokens, setNumberOfUniqueTerms, setRequest, setTermFrequency, stirlingPower

Methods inherited from class java.lang.Object

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

Field Detail

logger

protected static final org.apache.log4j.Logger logger

basicModel

protected BasicModel basicModel

The applied basic model for randomness.

afterEffect

protected AfterEffect afterEffect

The applied model for after effect (aka. first normalisation).

normalisation

protected Normalisation normalisation

The applied frequency normalisation method.

NORMALISATION_PREFIX

protected final java.lang.String NORMALISATION_PREFIX

The prefix of the package of the frequency normalisation methods.

See Also:

Constant Field Values

BASICMODEL_PREFIX

protected final java.lang.String BASICMODEL_PREFIX

The prefix of the package of the basic models for randomness.

See Also:

Constant Field Values

AFTEREFFECT_PREFIX

protected final java.lang.String AFTEREFFECT_PREFIX

The prefix of the package of the first normalisation methods by after effect.

See Also:

Constant Field Values

parameter

protected double parameter

The parameter of the frequency normalisation component.

ENABLE_NORMALISATION

protected boolean ENABLE_NORMALISATION

A boolean that indicates if the frequency normalisation is enabled.

ENABLE_AFTEREFFECT

protected boolean ENABLE_AFTEREFFECT

A boolean that indicates if the first normalisation by after effect is enabled.

Constructor Detail

DFRWeightingModel

public DFRWeightingModel(java.lang.String[] components)

The default constructor. Takes an array of strings to define the Basic Model, the After Effect component and the Normalisation component. If the array is less than 3 items in length, then empty strings will be passed instead of the After Effect and/or Normalisation components.

Parameters:

components - Corresponds to the names of the 3 DFR weighting models component names, as passed to initialise().

Method Detail

initialise

protected void initialise(java.lang.String basicModelName,
                          java.lang.String afterEffectName,
                          java.lang.String normalisationName)

Initialise the components in the DFR model. For each component, if a package is not specified, then a prefix will be applied. These are BASICMODEL_PREFIX, AFTEREFFECT_PREFIX and NORMALISATION_PREFIX respectively. Note that NORMALISATION_PREFIX includes a partial class name.

Parameters:

basicModelName - The name of the applied basic model for randomness. This component must be specified and can NOT be an empty string.

afterEffectName - The name of the applied first normalisation by after effect. An empty string to disable this component.

normalisationName - The name of the applied frequency normalisation component. An empty string to disable this component.

initialise

protected void initialise(java.lang.String basicModelName,
                          java.lang.String afterEffectName,
                          java.lang.String normalisationName,
                          double _parameter)

Initialise the components in the DFR model. For each component, if a package is not specified, then a prefix will be applied. These are BASICMODEL_PREFIX, AFTEREFFECT_PREFIX and NORMALISATION_PREFIX respectively. Note that NORMALISATION_PREFIX includes a partial class name.

Parameters:

basicModelName - The name of the applied basic model for randomness. This component must be specified and can NOT be an empty string.

afterEffectName - The name of the applied first normalisation by after effect. An empty string to disable this component.

normalisationName - The name of the applied frequency normalisation component. An empty string to disable this component.

_parameter - The applied parameter value of the frequency normalisation.

setParameter

public void setParameter(double value)

Set the frequency normalisation parameter.

Specified by:

setParameter in interface Model

Overrides:

setParameter in class WeightingModel

Parameters:

value - The given parameter value.

getParameter

public double getParameter()

Return the parameter set by setParameter()

Specified by:

getParameter in interface Model

Overrides:

getParameter in class WeightingModel

Returns:

parameter double value

getInfo

public final java.lang.String getInfo()

Returns the name of the model.

Specified by:

getInfo in interface Model

Specified by:

getInfo in class WeightingModel

Returns:

The name of the model.

score

public final double score(double tf,
                          double docLength)

Compute a weight for a term in a document.

Specified by:

score in class WeightingModel

Parameters:

tf - The term frequency in the document

docLength - the document's length

Returns:

the score assigned to a document with the given tf and docLength, and other preset parameters

score

public final double score(double tf,
                          double docLength,
                          double documentFrequency,
                          double termFrequency,
                          double queryTermWeight)

Compute a weight for a term in a document.

Specified by:

score in class WeightingModel

Parameters:

tf - The term frequency in the document

docLength - the document's length

documentFrequency - The document frequency of the term

termFrequency - the term frequency in the collection

queryTermWeight - the term frequency in the query

Returns:

the score assigned by the weighting model PL2.

setCollectionStatistics

public void setCollectionStatistics(CollectionStatistics _cs)

Description copied from class: WeightingModel

Sets collection statistics

Overrides:

setCollectionStatistics in class WeightingModel

setEntryStatistics

public void setEntryStatistics(EntryStatistics _es)

Description copied from class: WeightingModel

Sets entry statistics.

Overrides:

setEntryStatistics in class WeightingModel