es.uam.eps.ir.ranksys.diversity.distance.reranking

Class MMR<U,I>

java.lang.Object

es.uam.eps.ir.ranksys.novdiv.reranking.PermutationReranker<U,I>

es.uam.eps.ir.ranksys.novdiv.reranking.GreedyReranker<U,I>

es.uam.eps.ir.ranksys.novdiv.reranking.LambdaReranker<U,I>

es.uam.eps.ir.ranksys.diversity.distance.reranking.MMR<U,I>

Type Parameters:

U - type of the users

I - type of the items

All Implemented Interfaces:

Reranker<U,I>

public class MMR<U,I>
extends LambdaReranker<U,I>

Maximum marginal relevance re-ranker. C.-N. Ziegler, S.M. McNee, J.A. Konstan and G. Lausen. Improving recommendation lists through topic diversification. WWW 2005. J. Carbonell and J. Goldstein. The use of MMR, diversity-based reranking for reordering documents and producing summaries. SIGIR 1998.

Author:

Saúl Vargas (saul.vargas@uam.es), Pablo Castells (pablo.castells@uam.es)

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
class	MMR.UserMMR User re-ranker for MMR.

Nested classes/interfaces inherited from class es.uam.eps.ir.ranksys.novdiv.reranking.LambdaReranker

LambdaReranker.LambdaUserReranker

Nested classes/interfaces inherited from class es.uam.eps.ir.ranksys.novdiv.reranking.GreedyReranker

GreedyReranker.GreedyUserReranker<U,I>

Field Summary

Fields inherited from class es.uam.eps.ir.ranksys.novdiv.reranking.LambdaReranker

lambda

Fields inherited from class es.uam.eps.ir.ranksys.novdiv.reranking.GreedyReranker

cutoff

Constructor Summary

Constructors

Constructor and Description
MMR(double lambda, int cutoff, ItemDistanceModel<I> dist) Constructor.

Method Summary

Modifier and Type	Method and Description
protected LambdaReranker.LambdaUserReranker	getUserReranker(Recommendation<U,I> recommendation, int maxLength) Returns an instance of GreedyReranker.GreedyUserReranker that does the greedy selection.

Methods inherited from class es.uam.eps.ir.ranksys.novdiv.reranking.LambdaReranker

rerankPermutation

Methods inherited from class es.uam.eps.ir.ranksys.novdiv.reranking.PermutationReranker

getBasePerm, permuteRecommendation, rerankRecommendation

Methods inherited from class java.lang.Object

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

Constructor Detail

MMR

public MMR(double lambda,
           int cutoff,
           ItemDistanceModel<I> dist)

Constructor.

Parameters:

lambda - trade-off parameter of the linear combination

cutoff - how many items are re-ranked by the greedy selection.

dist - item distance model

Method Detail

getUserReranker

protected LambdaReranker.LambdaUserReranker getUserReranker(Recommendation<U,I> recommendation,
                                                            int maxLength)

Description copied from class: GreedyReranker

Returns an instance of GreedyReranker.GreedyUserReranker that does the greedy selection.

Specified by:

getUserReranker in class LambdaReranker<U,I>

Parameters:

recommendation - input recommendation to be re-ranked

maxLength - maximum length of the resulting re-ranked recommendation

Returns:

the GreedyReranker.GreedyUserReranker that does the re-ranking