uk.ac.gla.dcs.renaissance.ir.rf
Class MG4JRelevanceFeedback

java.lang.Object
  extended by uk.ac.gla.dcs.renaissance.ir.rf.MG4JRelevanceFeedback

public class MG4JRelevanceFeedback
extends Object

Relevance Feedback handling with MG4J (translates back and forth between internal and external ID)

Author:
B. Piwowarski

Nested Class Summary
static class MG4JRelevanceFeedback.MG4JDocument
          A document contains both the docno (external) & docid (internal) information.
static class MG4JRelevanceFeedback.MG4JFactory
           
 
Constructor Summary
MG4JRelevanceFeedback()
           
 
Method Summary
static Collection<bpiwowar.utils.Pair<Integer,Float>> get(it.unimi.dsi.mg4j.document.DocumentCollection collection, RelevanceFeedbackMethod method, String topicId, it.unimi.dsi.fastutil.objects.ObjectArrayList<it.unimi.dsi.mg4j.search.score.DocumentScoreInfo<it.unimi.dsi.fastutil.objects.Reference2ObjectMap<it.unimi.dsi.mg4j.index.Index,it.unimi.dsi.mg4j.query.SelectedInterval[]>>> results)
          Get relevance feedback using MG4J.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

MG4JRelevanceFeedback

public MG4JRelevanceFeedback()
Method Detail

get

public static final Collection<bpiwowar.utils.Pair<Integer,Float>> get(it.unimi.dsi.mg4j.document.DocumentCollection collection,
                                                                       RelevanceFeedbackMethod method,
                                                                       String topicId,
                                                                       it.unimi.dsi.fastutil.objects.ObjectArrayList<it.unimi.dsi.mg4j.search.score.DocumentScoreInfo<it.unimi.dsi.fastutil.objects.Reference2ObjectMap<it.unimi.dsi.mg4j.index.Index,it.unimi.dsi.mg4j.query.SelectedInterval[]>>> results)
Get relevance feedback using MG4J. This methods turns the output of a scorer into a collection of documents returned after relevance feedback. It supports different relevance feedback strategies.

Parameters:
collection - the document collection
method - the relevance feedback strategy
topicId - the topic ID
results - the results as returned by the base scorer
Returns:
the documents selected applying the relevance feedback strategy


Copyright © 2011. All Rights Reserved.