Terrier IR Platform
2.2.1

uk.ac.gla.terrier.matching.models.aftereffect
Class AfterEffect

java.lang.Object
  extended by uk.ac.gla.terrier.matching.models.aftereffect.AfterEffect
All Implemented Interfaces:
java.io.Serializable
Direct Known Subclasses:
B, L, LL

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

This class provides a contract for implementing the first normalisation by after effect models for the DFR framework. This is referred to as the component (1-prob2) in the DFR framework. Classes implementing this interface are used by the DFRWeightingModel.

Version:
$Revision: 1.11 $
Author:
Ben He
See Also:
DFRWeightingModel, Serialized Form

Constructor Summary
AfterEffect()
          A default constructor
 
Method Summary
abstract  double gain(double tf, double documentFrequency, double termFrequency)
          This method provides the contract for implementing first normalisation by after effect.
abstract  java.lang.String getInfo()
          Returns the name of the model.
 double getParameter()
           
 void setAverageDocumentLength(double value)
          Set the average document length, which is used for computing the prior for the first normalisation.
 void setParameter(double parameter)
           
 
Methods inherited from class java.lang.Object
equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

AfterEffect

public AfterEffect()
A default constructor

Method Detail

setAverageDocumentLength

public void setAverageDocumentLength(double value)
Set the average document length, which is used for computing the prior for the first normalisation.

Parameters:
value - The average document length.

getParameter

public double getParameter()
Returns:
the term frequency normalisation parameter

setParameter

public void setParameter(double parameter)
Parameters:
parameter - the term frequency normalisation parameter value to set

getInfo

public abstract java.lang.String getInfo()
Returns the name of the model.

Returns:
java.lang.String

gain

public abstract double gain(double tf,
                            double documentFrequency,
                            double termFrequency)
This method provides the contract for implementing first normalisation by after effect.

Parameters:
tf - The term frequency in the document
documentFrequency - The document frequency of the given query term
termFrequency - The frequency of the given term in the whole collection.
Returns:
The gain of having one more occurrence of the query term.

Terrier IR Platform
2.2.1

Terrier Information Retrieval Platform 2.2.1. Copyright 2004-2008 University of Glasgow