es.ucm.fdi.gaia.recolibry.implementations.mahout.cfItem

Class RecommenderMahoutCFItem

java.lang.Object

es.ucm.fdi.gaia.recolibry.implementations.mahout.cfItem.RecommenderMahoutCFItem

All Implemented Interfaces:

RecommenderAlgorithm

public class RecommenderMahoutCFItem
extends Object
implements RecommenderAlgorithm

Implementation of a generic item recommender system using Mahout framework.

Version:

1.0

Author:

Jose L. Jorro-Aragoneses

Constructor Summary

Constructors

Constructor and Description
RecommenderMahoutCFItem(DataModelFactory dataModel, org.apache.mahout.cf.taste.similarity.ItemSimilarity itemSimilarity, int numResults) Method to build the recommender system.

Method Summary

Modifier and Type	Method and Description
boolean	close() Method to finish and close a recommender algorithm.
boolean	init() Method to initialize a recommender algorithm.
List<RecommenderResult>	recommend(Query query) Method to calculate a recommendation in the algorithm implemented.

Methods inherited from class java.lang.Object

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

Constructor Detail

RecommenderMahoutCFItem

@Inject
public RecommenderMahoutCFItem(DataModelFactory dataModel,
                                       org.apache.mahout.cf.taste.similarity.ItemSimilarity itemSimilarity,
                                       int numResults)

Method to build the recommender system.

Parameters:

dataModel - data model that saves all information used by the recommender system.

itemSimilarity - similarity function configure to the recommender system.

numResults - maximum number of results returned by the recommender system.

Method Detail

init

public boolean init()

Description copied from interface: RecommenderAlgorithm

Method to initialize a recommender algorithm. For example, if the algorithm has a training step, it is implemented in this method.

Specified by:

init in interface RecommenderAlgorithm

Returns:

true when the algorithm is initialized correctly, false in other case.

recommend

public List<RecommenderResult> recommend(Query query)

Description copied from interface: RecommenderAlgorithm

Method to calculate a recommendation in the algorithm implemented.

Specified by:

recommend in interface RecommenderAlgorithm

Parameters:

query - Query that contains the information to make a recomendation.

Returns:

List of elements recommended by the algorithm.

close

public boolean close()

Description copied from interface: RecommenderAlgorithm

Method to finish and close a recommender algorithm. For example, if the algorithm needs to save a new model, it is implemented in this method.

Specified by:

close in interface RecommenderAlgorithm

Returns:

true when the algorithm is finished and closed correctly, false in other case.