net.librec.recommender.cf.ranking

Class GBPRRecommender

java.lang.Object

net.librec.recommender.AbstractRecommender

net.librec.recommender.MatrixFactorizationRecommender

net.librec.recommender.cf.ranking.GBPRRecommender

All Implemented Interfaces:

Recommender

@ModelData(value={"isRanking","gbpr","userFactors","itemFactors","trainMatrix"})
public class GBPRRecommender
extends MatrixFactorizationRecommender

Pan and Chen, GBPR: Group Preference Based Bayesian Personalized Ranking for One-Class Collaborative Filtering, IJCAI 2013.

Author:

guoguibing and Keqiang Wang

Field Summary

Fields

Modifier and Type	Field and Description
protected static java.lang.String	cacheSpec Guava cache configuration
protected com.google.common.cache.LoadingCache<java.lang.Integer,java.util.List<java.lang.Integer>>	itemUsersCache user-items cache, item-users cache
protected double	regBias bias regularization
protected com.google.common.cache.LoadingCache<java.lang.Integer,java.util.List<java.lang.Integer>>	userItemsCache user-items cache, item-users cache

Fields inherited from class net.librec.recommender.MatrixFactorizationRecommender

initMean, initStd, itemFactors, learnRate, maxLearnRate, numFactors, numIterations, regItem, regUser, userFactors

Fields inherited from class net.librec.recommender.AbstractRecommender

conf, context, decay, earlyStop, globalMean, isBoldDriver, isRanking, itemMappingData, lastLoss, LOG, loss, maxRate, minRate, numItems, numRates, numUsers, ratingScale, recommendedList, testMatrix, topN, trainMatrix, userMappingData, validMatrix, verbose

Constructor Summary

Constructors

Constructor and Description
GBPRRecommender()

Method Summary

Methods

Modifier and Type	Method and Description
protected double	predict(int userIdx, int itemIdx) predict a specific rating for user userIdx on item itemIdx.
protected double	predict(int userIdx, int itemIdx, java.util.Set<java.lang.Integer> groupSet)
protected void	setup() setup init member method
protected void	trainModel() train Model

Methods inherited from class net.librec.recommender.MatrixFactorizationRecommender

updateLRate

Methods inherited from class net.librec.recommender.AbstractRecommender

cleanup, evaluate, evaluateMap, getContext, getDataModel, getRecommendedList, isConverged, loadModel, predict, recommend, recommend, recommendRank, recommendRating, saveModel, setContext

Methods inherited from class java.lang.Object

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

Field Detail

regBias

protected double regBias

bias regularization

userItemsCache

protected com.google.common.cache.LoadingCache<java.lang.Integer,java.util.List<java.lang.Integer>> userItemsCache

user-items cache, item-users cache

itemUsersCache

protected com.google.common.cache.LoadingCache<java.lang.Integer,java.util.List<java.lang.Integer>> itemUsersCache

user-items cache, item-users cache

cacheSpec

protected static java.lang.String cacheSpec

Guava cache configuration

Constructor Detail

GBPRRecommender

public GBPRRecommender()

Method Detail

setup

protected void setup()
              throws LibrecException

Description copied from class: MatrixFactorizationRecommender

setup init member method

Overrides:

setup in class MatrixFactorizationRecommender

Throws:

LibrecException - if error occurs during setting up

trainModel

protected void trainModel()
                   throws LibrecException

Description copied from class: AbstractRecommender

train Model

Specified by:

trainModel in class AbstractRecommender

Throws:

LibrecException - if error occurs during training model

predict

protected double predict(int userIdx,
             int itemIdx,
             java.util.Set<java.lang.Integer> groupSet)
                  throws LibrecException

Throws:

LibrecException

predict

protected double predict(int userIdx,
             int itemIdx)

Description copied from class: MatrixFactorizationRecommender

predict a specific rating for user userIdx on item itemIdx.

Overrides:

predict in class MatrixFactorizationRecommender

Parameters:

userIdx - user index

itemIdx - item index

Returns:

predictive rating for user userIdx on item itemIdx with bound