User Tools

Site Tools


algorithmlist

# Algorithm list

Recommender Algorithm List

superClass directory path short name algorithm
MatrixRecommender baseline constantguess ConstantGuessRecommender
MatrixRecommender baseline globalaverage GlobalAverageRecommender
MatrixRecommender baseline itemaverage ItemAverageRecommender
MatrixProbabilisticGraphicalRecommender baseline itemcluster ItemClusterRecommender
MatrixRecommender baseline mostpopular MostPopularRecommender
MatrixRecommender baseline randomguess RandomGuessRecommender
MatrixRecommender baseline useraverage UserAverageRecommender
MatrixProbabilisticGraphicalRecommender baseline usercluster UserClusterRecommender
MatrixProbabilisticGraphicalRecommender cf bhfree BHFreeRecommender
MatrixProbabilisticGraphicalRecommender cf bucm BUCMRecommender
MatrixRecommender cf itemknn ItemKNNRecommender
MatrixRecommender cf itemknn ItemKNNRecommender
MatrixRecommender cf userknn UserKNNRecommender
MatrixRecommender cf userknn UserKNNRecommender
MatrixFactorizationRecommender cf.ranking aobpr AoBPRRecommender
MatrixProbabilisticGraphicalRecommender cf.ranking aspectmodelranking AspectModelRecommender
MatrixFactorizationRecommender cf.ranking bnppf BNPPFRecommeder
MatrixFactorizationRecommender cf.ranking bpoissmf BPoissMFRecommender
MatrixFactorizationRecommender cf.ranking bpr BPRRecommender
MatrixFactorizationRecommender cf.ranking climf CLIMFRecommender
MatrixFactorizationRecommender cf.ranking cofiset CoFiSetRecommender
MatrixFactorizationRecommender cf.ranking eals EALSRecommender
MatrixFactorizationRecommender cf.ranking fismauc FISMaucRecommender
MatrixFactorizationRecommender cf.ranking fismrmse FISMrmseRecommender
MatrixFactorizationRecommender cf.ranking gbpr GBPRRecommender
MatrixProbabilisticGraphicalRecommender cf.ranking itembigram ItemBigramRecommender
MatrixProbabilisticGraphicalRecommender cf.ranking lda LDARecommender
MatrixFactorizationRecommender cf.ranking listrankmf ListRankMFRecommender
MatrixFactorizationRecommender cf.ranking nmfitemitem NMFItemItemRecommender
MatrixProbabilisticGraphicalRecommender cf.ranking plsa PLSARecommender
MatrixFactorizationRecommender cf.ranking pnmf PNMFRecommender
MatrixFactorizationRecommender cf.ranking rankals RankALSRecommender
MatrixFactorizationRecommender cf.ranking rankpmf RankPMFRecommender
MatrixFactorizationRecommender cf.ranking ranksgd RankSGDRecommender
MatrixFactorizationRecommender cf.ranking slim SLIMRecommender
MatrixFactorizationRecommender cf.ranking wbpr WBPRRecommender
MatrixFactorizationRecommender cf.ranking wrmf WRMFRecommender
MatrixProbabilisticGraphicalRecommender cf.rating aspectmodelrating AspectModelRecommender
cf.rating.BiasedMFRecommender cf.rating asvdpp ASVDPlusPlusRecommender
MatrixFactorizationRecommender cf.rating biasedmf BiasedMFRecommender
MatrixFactorizationRecommender cf.rating bpmf BPMFRecommender
FactorizationMachineRecommender cf.rating ffm FFMRecommender
FactorizationMachineRecommender cf.rating fmals FMALSRecommender
FactorizationMachineRecommender cf.rating fmftrl FMFTRLRecommender
FactorizationMachineRecommender cf.rating fmsgd FMSGDRecommender
MatrixProbabilisticGraphicalRecommender cf.rating gplsa GPLSARecommender
MatrixFactorizationRecommender cf.rating irrg IRRGRecommender
MatrixProbabilisticGraphicalRecommender cf.rating ldcc LDCCRecommender
MatrixFactorizationRecommender cf.rating llorma LLORMARecommender
MatrixFactorizationRecommender cf.rating mfals MFALSRecommender
MatrixFactorizationRecommender cf.rating nmf NMFRecommender
MatrixFactorizationRecommender cf.rating pmf PMFRecommender
MatrixRecommender cf.rating rbm RBMRecommender
MatrixFactorizationRecommender cf.rating remf ReMFRecommender
MatrixFactorizationRecommender cf.rating rfrec RFRecRecommender
cf.rating.BiasedMFRecommender cf.rating svdpp SVDPlusPlusRecommender
MatrixProbabilisticGraphicalRecommender cf.rating urp URPRecommender
TensorRecommender content convmf ConvMFRecommender
TensorRecommender content efm EFMRecommender
TensorRecommender content hft HFTRecommender
TensorRecommender content tfidf TFIDFRecommender
TensorRecommender content topicmfat TopicMFATRecommender
TensorRecommender content topicmfmt TopicMFMTRecommender
FactorizationMachineRecommender context.ranking dlambdafm DLambdaFMRecommender
SocialRecommender context.ranking sbpr SBPRRecommender
TensorRecommender context.rating cptf CPTFRecommender
SocialRecommender context.rating rste RSTERecommender
SocialRecommender context.rating socialmf SocialMFRecommender
SocialRecommender context.rating sorec SoRecRecommender
SocialRecommender context.rating soreg SoRegRecommender
cf.rating.BiasedMFRecommender context.rating timesvd TimeSVDRecommender
SocialRecommender context.rating trustmf TrustMFRecommender
SocialRecommender context.rating trustsvd TrustSVDRecommender
MatrixFactorizationRecommender cf.ranking fismauc FISMaucRecommender
MatrixFactorizationRecommender cf.ranking fismrmse FISMrmseRecommender
MatrixFactorizationRecommender cf.rating biasedmf BiasedMFRecommender
MatrixFactorizationRecommender cf.rating biasedmf BiasedMFRecommender
MatrixFactorizationRecommender cf.rating nmf NMFRecommender
MatrixFactorizationRecommender cf.rating pmf PMFRecommender
MatrixRecommender ext associationrule AssociationRuleRecommender
MatrixRecommender ext bipolarslopeone BipolarSlopeOneRecommender
MatrixRecommender ext external ExternalRecommender
MatrixRecommender ext personalitydiagnosis PersonalityDiagnosisRecommender
cf.ranking.RankSGDRecommender ext prankd PRankDRecommender
MatrixRecommender ext slopeone SlopeOneRecommender
MatrixRecommender hybrid hybrid HybridRecommender
MatrixRecommender cf itemknn ItemKNNRecommender
MatrixFactorizationRecommender cf.rating pmf PMFRecommender
MatrixRecommender cf userknn UserKNNRecommender
MatrixRecommender nn.ranking cdae CDAERecommender
MatrixRecommender nn.rating autorec AutoRecRecommender
MatrixFactorizationRecommender poi rankgeofm RankGeoFMRecommender
AbstractRecommender poi usg USGRecommender

Algorithm Configuration List

Baseline

ConstantGuessRecommender
rec.recommender.class=constantguess
GlobalAverageRecommender
rec.recommender.class=globalaverage
ItemAverageRecommender
rec.recommender.class=itemaverage
ItemClusterRecommender
rec.recommender.class=itemcluster
rec.pgm.number=10
rec.iterator.maximum=20
MostPopularRecommender
rec.recommender.class=mostpopular
rec.recommender.isranking=true
RandomGuessRecommender
rec.recommender.class=randomguess
# setting dataset format(UIR, UIRT)
data.cache = false
UserAverageRecommender
rec.recommender.class=useraverage
UserClusterRecommender
rec.recommender.class=usercluster
rec.factory.number=10
rec.iterator.maximum=20

Collaborative Filtering (item ranking)

AoBPRRecommender
rec.recommender.class=aobpr
rec.item.distribution.parameter = 500
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
AspectModelRecommender
rec.recommender.class=aspectmodelranking
rec.iterator.maximum=20
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
data.splitter.cv.number=5
rec.pgm.burnin=10
rec.pgm.samplelag=10
rec.topic.number=10
BNPPFRecommeder
rec.recommender.class=bnppf
rec.iterator.maximum=1
rec.factor.number=10
rec.recommender.isranking=true
rec.recommender.ranking.topn=10

rec.alpha=0.3
rec.c=0.3
rec.a=0.3
rec.b=0.3
BPoissMFRecommender
rec.recommender.class=bpoissmf
rec.iterator.maximum=100
rec.factor.number=10
rec.recommender.isranking=true
rec.recommender.ranking.topn=10

data.convert.binarize.threshold=0.0

rec.a=0.3
rec.a.prime=0.3
rec.b.prime=1.0
rec.c=0.3
rec.c.prime=0.3
rec.d.prime=1.0
BPRRecommender
rec.recommender.class=bpr
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=50
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learnRate.bolddriver=false
rec.learnRate.decay=1.0
data.convert.binarize.threshold=0.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10

CLIMFRecommender
rec.recommender.class=climf
rec.iterator.learnrate=0.001
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10

CoFiSetRecommender
rec.recommender.class=cofiset
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=20
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.bias.regularization=0.01
rec.factor.number=10
rec.presence.size=2
rec.absence.size=1
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
EALSRecommender
rec.recommender.class=eals
rec.iterator.maximum=20
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=200
rec.recommender.isranking=true
rec.recommender.ranking.topn=10

#0:eALS MF; 1:WRMF; 2: both
rec.eals.wrmf.judge=1

#the overall weight of missing data c0
rec.eals.overall=128

#the significance level of popular items over un-popular ones
rec.eals.ratio=0.4

#confidence weight coefficient, alpha in original paper
rec.wrmf.weight.coefficient=1.0

FISMaucRecommender
rec.recommender.class=fismauc
rec.iteration.learnrate=0.00001
rec.iterator.maximum=5
rec.recommender.isranking=true

rec.recommender.rho=0.5
rec.recommender.beta=0.6
rec.recommender.alpha=0.9
rec.recommender.gamma=0.1
rec.factor.number=10

guava.cache.spec=maximumSize=200,expireAfterAccess=2m
FISMrmseRecommender
rec.recommender.class=fismrmse
rec.iteration.learnrate=0.0001
rec.iterator.maximum=14
rec.recommender.isranking=true

rec.recommender.rho=1
rec.recommender.beta=0.6
rec.recommender.alpha=0.8
rec.factor.number=10
rec.recommender.userBiasReg=0.1
rec.recommender.itemBiasReg=0.1

guava.cache.spec=maximumSize=200,expireAfterAccess=2m

GBPRRecommender
rec.recommender.class=gbpr
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
rec.gpbr.rho=1.5
rec.gpbr.gsize=2

ItemBigramRecommender
rec.recommender.class=itembigram
data.column.format=UIRT
data.input.path=test/datamodeltest/ratings-date.txt
rec.iterator.maximum=100
rec.topic.number=10
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
rec.user.dirichlet.prior=0.01
rec.topic.dirichlet.prior=0.01
rec.pgm.burnin=10
rec.pgm.samplelag=10

LDARecommender
rec.recommender.class=lda
rec.iterator.maximum=1000
rec.topic.number = 10
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
rec.user.dirichlet.prior=0.01
rec.topic.dirichlet.prior=0.01
rec.pgm.burnin=100
rec.pgm.samplelag=10
data.splitter.cv.number=5
# (0.0 maybe a better choose than -1.0)
data.convert.binarize.threshold=0.0

ListRankMFRecommender
rec.recommender.class=listrankmf
data.splitter.ratio=user
rec.iterator.learnrate=1.0
rec.iterator.learnrate.maximum=100
rec.iterator.maximum=30
rec.user.regularization=0.06
rec.item.regularization=0.06
rec.factor.number=5
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
NMFItemItemRecommender
rec.recommender.class=nmfitemitem
rec.iterator.maximum=50
rec.factor.number=20
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
rec.nmfitemitem.do_not_estimate_yourself=true
rec.nmfitemitem.adaptive_update_rules=true
rec.nmfitemitem.parallelize_split_user_size=5000
PLSARecommender
rec.recommender.class=plsa
rec.iteration.learnrate=0.01
rec.iterator.maximum=100
rec.recommender.isranking=true
rec.topic.number = 10
rec.recommender.ranking.topn=10
# (0.0 maybe a better choose than -1.0)
data.convert.binarize.threshold=0.0
PNMFRecommender
rec.recommender.class=pnmf
rec.iterator.maximum=25
rec.factor.number=10
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
data.convert.binarize.threshold=0
RankALSRecommender
rec.recommender.class=rankals
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10

rec.rankals.support.weight=true
RankPMFRecommender
rec.recommender.class=rankpmf
rec.iterator.maximum=20
rec.confidence.a = 1
rec.confidence.b = 0.01
rec.user.regularization=0.1
rec.item.regularization=10
rec.factor.number=10
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
RankSGDRecommender
rec.recommender.class=ranksgd
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=30
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
SLIMRecommender
rec.recommender.class=slim

rec.similarity.class=cos
#can only use item similarity
rec.recommender.similarities=item
rec.iterator.maximum=40
rec.similarity.shrinkage=10
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
rec.neighbors.knn.number=50
rec.recommender.earlystop=true

rec.slim.regularization.l1=1
rec.slim.regularization.l2=5

WBPRRecommender
rec.recommender.class=wbpr
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=20
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.bias.regularization=0.01
rec.factor.number=128
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
WRMFRecommender
rec.recommender.class=wrmf
rec.iterator.maximum=20
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=20
rec.recommender.isranking=true
rec.recommender.ranking.topn=10

#confidence weight coefficient, alpha in original paper
rec.wrmf.weight.coefficient=4.0

Collaborative Filtering (rating prediction)

AspectModelRecommender
rec.recommender.class=aspectmodelrating
rec.iteration.learnrate=0.01
rec.iterator.maximum=100
ASVDPlusPlusRecommender
rec.recommender.class=asvdpp
rec.iteration.learnrate=0.01
rec.iterator.maximum=20
BiasedMFRecommender
rec.recommender.class=biasedmf
rec.iterator.learnrate=0.002
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.bias.regularization=0.01
rec.factor.number=20
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
BPMFRecommender
rec.recommender.class=bpmf
rec.iterator.maximum=150
rec.factor.number=20

rec.recommender.user.mu=0.0
rec.recommender.item.mu=0.0

rec.recommender.user.beta=1.0
rec.recommender.item.beta=1.0

rec.recommender.user.wishart.scale=1.0
rec.recommender.item.wishart.scale=1.0

rec.recommender.rating.sigma=2.0

# rec.learnrate.bolddriver=false
# rec.learnrate.decay=1.0
rec.recommender.gibbs.iterations = 1
FFMRecommender
data.input.path=test/datamodeltest/ratings.arff
data.column.format=UIR
data.model.splitter=ratio
data.convertor.format=arff
data.model.format=arff

rec.recommender.class=ffm
rec.iterator.learnRate=0.001
rec.iterator.maximum=100
rec.factor.number=10
FMALSRecommender
data.input.path=test/datamodeltest/ratings.arff
data.column.format=UIR
data.model.splitter=ratio
data.convertor.format=arff
data.model.format=arff

rec.recommender.class=fmals
rec.iterator.learnRate=0.01
rec.iterator.maximum=100
rec.factor.number=10
FMFTRLRecommender
data.input.path=test/datamodeltest/ratings.arff
data.column.format=UIR
data.model.splitter=ratio
data.convertor.format=arff
data.model.format=arff

rec.recommender.class=fmftrl

rec.iterator.maximum=30
rec.factor.number=10

rec.regularization.lambda1=0.05
rec.regularization.lambda2=1.0
rec.learningRate.alpha=0.015
rec.learningRate.beta=1
FMSGDRecommender
data.input.path=test/datamodeltest/ratings.arff
data.column.format=UIR
data.model.splitter=ratio
data.convertor.format=arff
data.model.format=arff

rec.recommender.class=fmsgd
rec.iterator.learnRate=0.001
rec.iterator.maximum=100
rec.factor.number=10
GPLSARecommender
rec.recommender.class=gplsa
rec.iteration.learnrate=0.01
rec.iterator.maximum=100
rec.recommender.smoothWeight=2
rec.recommender.isranking=false
rec.topic.number = 10
IRRGRecommender
rec.recommender.class=irrg
rec.iterator.learnrate=0.001
rec.iterator.learnrate.maximum=10

rec.iterator.maximum=200
rec.user.regularization=0.001
rec.item.regularization=0.001
rec.alpha=0.1
rec.factor.number=10
rec.learnrate.bolddriver=true
rec.learnrate.decay=1.0

LDCCRecommender
rec.recommender.class=ldcc
rec.iteration.learnrate=0.01
rec.iterator.maximum=1000
rec.pgm.burnin=100
rec.pgm.samplelag=10

rec.pgm.number.users=10
rec.pgm.number.items=10
rec.pgm.user.alpha=0.1
rec.pgm.item.alpha=0.1
rec.pgm.rating.beta=0.2
LLORMARecommender
rec.recommender.class=llorma
rec.global.factors.num=10
rec.global.iteration.learnrate=0.0005
rec.global.user.regularization=0.1
rec.global.item.regularization=0.1
rec.global.iteration.maximum=200
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=200
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=6
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.model.num=55
rec.thread.count=8
MFALSRecommender
rec.recommender.class=mfals
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
NMFRecommender
rec.recommender.class=nmf
rec.iterator.maximum=10
rec.factor.number=100
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
PMFRecommender
rec.recommender.class=pmf
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=70
rec.user.regularization=0.08
rec.item.regularization=0.08
rec.factor.number=6
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
RBMRecommender
rec.recommender.class=rbm
rec.iterator.maximum=20
data.input.path=movielens/ml-100k/ratings.txt
rec.factor.number=500
rec.epsilonw=0.01
rec.epsilonvb=0.01
rec.epsilonhb=0.01
rec.tstep=1
rec.momentum=0.1
rec.lamtaw=0.01
rec.lamtab=0.0
rec.predictiontype=mean
ReMFRecommender
data.appender.class=auxiliary
data.appender.path=twitter/user_hierarchy.txt
data.input.path=twitter/london.txt

rec.recommender.class=remf
rec.iterator.learnrate=0.0001
rec.iterator.learnrate.maximum=10

rec.iterator.maximum=130
rec.user.regularization=0.05
rec.item.regularization=0.05
rec.alpha=0.01
rec.side=user
rec.factor.number=10
rec.learnrate.bolddriver=true
rec.learnrate.decay=1.0
RFRecRecommender
rec.recommender.class=rfrec
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
SVDPlusPlusRecommender
rec.recommender.class=svdpp
rec.iterator.learnrate=0.002
rec.iterator.learnrate.maximum=0.05
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.impItem.regularization=0.01
rec.bias.regularization=0.01
rec.factor.number=20
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
URPRecommender
rec.recommender.class=urp
rec.iteration.learnrate=0.01
rec.iterator.maximum=100
UserKNNRecommender
#data.input.path=filmtrust/rating
data.column.format=UIRT

rec.similarity.class=pcc
rec.neighbors.knn.number=20
rec.recommender.class=userknn
rec.recommender.similarities=user
rec.recommender.isranking=false
rec.recommender.ranking.topn=10
rec.filter.class=generic
rec.similarity.shrinkage=10

Collaborative Filtering (rating prediction and item ranking)

BHFreeRecommender
rec.recommender.class=bhfree
rec.pgm.burnin=10
rec.pgm.samplelag=10
rec.iterator.maximum=100
# true for item ranking, false for rating prediction
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
BUCMRecommender
rec.recommender.class=bucm
rec.pgm.burnin=10
rec.pgm.samplelag=10

rec.iterator.maximum=100
rec.pgm.topic.number=10
rec.bucm.alpha=0.01
rec.bucm.beta=0.01
rec.bucm.gamma=0.01
# true for item ranking, false for rating prediction
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
ItemKNNRecommender
rec.recommender.class=itemknn
# true for item ranking, false for rating prediction
rec.recommender.isranking=false
rec.recommender.ranking.topn=10
rec.recommender.similarities=item
rec.similarity.class=pcc
rec.neighbors.knn.number=50
rec.similarity.shrinkage=10
UserKNNRecommender
rec.similarity.class=pcc
rec.neighbors.knn.number=50
rec.recommender.class=userknn
rec.recommender.similarities=user
# true for item ranking, false for rating prediction
rec.recommender.isranking=false
rec.recommender.ranking.topn=10
rec.filter.class=generic
rec.similarity.shrinkage=10

Content

ConvMFRecommender
# the dataset needs decompressing first
data.input.path=test/hfttest/digital_music.arff
#data.input.path=test/hfttest/musical_instruments.arff
data.convertor.format=arff
data.model.format=arff
rec.recommender.class=convmf
rec.iterator.learnrate=0.001
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=20
rec.user.regularization=0.1
rec.item.regularization=0.1
rec.factor.number=32
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.word2vec.path=test/hfttest/word2vec_org
rec.word2vec.dimension=200
rec.document.length=100
rec.featuremap.num=32
EFMRecommender
# the dataset here is sampled from th Labeled DC reviews dataset at http://yongfeng.me/dataset/
# arrange your own dataset if you need.
data.input.path=test/efmtest/dc_dense.arff
data.splitter.trainset.ratio=0.8
data.convertor.format=arff
data.model.format=arff
rec.recommender.class=efm
rec.iterator.maximum=50
rec.factor.number=10
rec.factor.explicit=5
rec.regularization.lambdax=1
rec.regularization.lambday=1
rec.regularization.lambdau=0.01
rec.regularization.lambdah=0.01
rec.regularization.lambdav=0.01

rec.explain.flag=true
rec.explain.userids=480 8517 550
rec.explain.numfeature=5


HFTRecommender
# The training approach is SGD instead of L-BFGS, so it can be slow if the dataset
# is big. if you want a quick test, try the path : test/hfttest/musical_instruments.arff
# path of the full dataset is : test/hfttest/musical_instruments_full.arff
data.input.path=test/hfttest/musical_instruments.arff
data.convertor.format=arff
data.model.format=arff
rec.recommender.class=hft
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=2
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.lambda.user=0.05
rec.recommender.lambda.item=0.05
rec.bias.regularization = 0.01
TFIDFRecommender
# The training approach is SGD instead of L-BFGS, so it can be slow if the dataset
# is big. if you want a quick test, try the path : test/hfttest/musical_instruments.arff
# path of the full dataset is : test/hfttest/musical_instruments_full.arff
data.convert.sep = ,
data.input.path=test/hfttest/musical_instruments.arff
data.convertor.format=arff
data.model.format=arff
rec.recommender.class=tfidf
rec.recommender.isranking=true
#rec.iterator.learnrate=0.01
#rec.iterator.learnrate.maximum=0.01
#rec.iterator.maximum=2
#rec.user.regularization=0.01
#rec.item.regularization=0.01
#rec.factor.number=10
#rec.learnrate.bolddriver=false
#rec.learnrate.decay=1.0
#rec.recommender.lambda.user=0.05
#rec.recommender.lambda.item=0.05
#rec.bias.regularization = 0.01
TopicMFATRecommender
# The training approach is SGD instead of L-BFGS, so it can be slow if the dataset
# is big. if you want a quick test, try the path : test/hfttest/musical_instruments.arff
# path of the full dataset is : test/hfttest/musical_instruments_full.arff
data.input.path=test/hfttest/digital_music.arff
data.convertor.format=arff
data.model.format=arff
rec.recommender.class=topicmfat
rec.regularization.lambda=0.001
rec.regularization.lambdaU=0.001
rec.regularization.lambdaV=0.001
rec.regularization.lambdaB=0.001
rec.topic.number=10
rec.iterator.learnrate=0.01
rec.iterator.maximum=10
rec.init.mean=0.0
rec.init.std=0.01
TopicMFMTRecommender
# The training approach is SGD instead of L-BFGS, so it can be slow if the dataset
# is big. if you want a quick test, try the path : test/hfttest/musical_instruments.arff
# path of the full dataset is : test/hfttest/musical_instruments_full.arff
data.input.path=test/hfttest/digital_music.arff
data.convertor.format=arff
data.model.format=arff
rec.recommender.class=topicmfmt
rec.regularization.lambda=0.001
rec.regularization.lambdaU=0.001
rec.regularization.lambdaV=0.001
rec.regularization.lambdaB=0.001
rec.topic.number=10
rec.iterator.learnrate=0.01
rec.iterator.maximum=10
rec.init.mean=0.0
rec.init.std=0.01

Context(item ranking)

DLambdaFMRecommender
rec.recommender.class=dlambdafm

data.convertor.format=arff
data.model.format=arff
data.input.path=test/lambdafm/music.arff
data.convert.binarize.threshold=0.0

rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=30
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=30
rec.learnRate.bolddriver=false
rec.learnRate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
rec.recommender.rho=0.3
rec.recommender.lossf=2
SBPRRecommender
data.appender.class=social
data.appender.path=test/test-append-dir

rec.recommender.class=sbpr
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=200
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.social.regularization=0.01
rec.bias.regularization=0.01
rec.factor.number=128
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.earlystop=false
rec.recommender.verbose=true
rec.recommender.isranking=true
rec.recommender.ranking.topn=10

Context(rating prediction)

CPTFRecommender
data.input.path=test/datamodeltest/ratings.arff
data.column.format=UIR
data.model.splitter=ratio
data.convertor.format=arff
data.model.format=arff
rec.random.seed=1

rec.recommender.class=cptf
rec.iteration.learnrate=0.01
rec.iterator.maximum=100
rec.factor.number=20
rec.tensor.regularization=0.05

rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
RSTERecommender
data.appender.class=social
data.appender.path=test/test-append-dir

rec.recommender.class=rste
rec.iterator.learnrate=0.05
rec.iterator.learnrate.maximum=0.05
rec.iterator.maximum=200
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.social.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.earlystop=false
rec.recommender.verbose=true
rec.user.social.ratio=0.8
SocialMFRecommender
data.appender.class=social
data.appender.path=test/test-append-dir

rec.recommender.class=socialmf
rec.iterator.learnrate=0.05
rec.iterator.learnrate.maximum=0.05
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.social.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.earlystop=false
rec.recommender.verbose=true
SoRecRecommender
data.appender.class=social
data.appender.path=test/test-append-dir

rec.recommender.class=sorec
rec.iterator.learnrate=0.05
rec.iterator.learnrate.maximum=0.05
rec.iterator.maximum=1000
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.rate.social.regularization=0.01
rec.user.social.regularization=0.01
rec.social.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.earlystop=false
rec.recommender.verbose=true
SoRegRecommender
data.appender.class=social
data.appender.path=test/test-append-dir

rec.recommender.class=soreg
rec.recommender.similarities=social
rec.similarity.class=pcc
rec.iterator.learnrate=0.001
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=10
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.social.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.earlystop=false
rec.recommender.verbose=true
rec.similarity.shrinkage=10
TimeSVDRecommender
rec.recommender.class=timesvd
data.column.format=UIRT
data.input.path=test/ratings-date.txt
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=100
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.learnrate.decay=1.0
TrustMFRecommender
data.appender.class=social
data.appender.path=test/test-append-dir

rec.recommender.class=trustmf
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=30
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.social.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.earlystop=false
rec.recommender.verbose=true
rec.social.model=T
TrustSVDRecommender
data.appender.class=social
data.appender.path=test/test-append-dir

rec.recommender.class=trustsvd
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=50
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.social.regularization=0.01
rec.bias.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.earlystop=false
rec.recommender.verbose=true

Extra

AssociationRuleRecommender
rec.recommender.class=associationrule
ExternalRecommender
rec.recommender.class=external
PersonalityDiagnosisRecommender
rec.recommender.class=personalitydiagnosis
rec.PersonalityDiagnosis.sigma=0.1
PRankDRecommender
rec.recommender.class=prankd
rec.similarity.class=cos
rec.recommender.similarities=item
rec.similarity.shrinkage=10
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=50
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.sim.filter=4.0
SlopeOneRecommender
rec.recommender.class=slopeone
rec.eval.enable=true
rec.iterator.maximum=50
rec.factory.number=30
rec.iterator.learn.rate=0.001
rec.recommender.lambda.user=0.05
rec.recommender.lambda.item=0.05

Hybrid

HybridRecommender
rec.recommender.class=hybrid
rec.hybrid.lambda=0.1
rec.iterator.maximum=50
rec.factory.number=30
rec.iterator.learn.rate=0.001
rec.recommender.lambda.user=0.05
rec.recommender.lambda.item=0.05

Neural Network

CDAERecommender
rec.recommender.class=cdae
rec.iterator.learnrate=0.1
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=1000
rec.weight.regularization=0.01
rec.hidden.dimension=200
rec.hidden.activation=sigmoid
rec.output.activation=identity
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
data.convert.binarize.threshold=3
AutoRecRecommender
rec.recommender.class=autorec
rec.iterator.learnrate=0.01
rec.iterator.learnrate.maximum=0.01
rec.iterator.maximum=200
rec.weight.regularization=0.001
rec.hidden.dimension=200
rec.hidden.activation=sigmoid
rec.output.activation=identity
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0

Point of Interest

RankGeoFMRecommender
data.appender.class=location
data.appender.path=poi/FourSquare/FoursquareLocation.txt
data.input.path=poi/FourSquare/checkin/trainData.txt
data.model.splitter=testset
data.testset.path=poi/FourSquare/checkin/testData.txt
rec.recommender.class=rankgeofm
rec.factor.number=100
rec.iterator.learnrate=0.001
rec.iterator.learnrate.maximum=0.001
rec.iterator.maximum=200
rec.regularization.C=1.0
rec.regularization.alpha=0.2
rec.ranking.epsilon=0.3
rec.item.knn=300
rec.learnrate.bolddriver=false
rec.learnrate.decay=1.0
rec.recommender.isranking=true
rec.eval.enable=true
rec.recommender.ranking.topn=10
USGRecommender
data.appender.class=location
data.appender.path=poi/Gowalla/Gowalla_poi_coos.txt
data.input.path=poi/Gowalla/checkin/Gowalla_train.txt
data.model.splitter=testset
##all user test set
#data.testset.path=poi/Gowalla/checkin/Gowalla_test.txt
##small user test set
data.testset.path=poi/Gowalla/checkin/testDataFor101users.txt
data.social.path=poi/Gowalla/Gowalla_social_relations.txt
data.convert.binarize.threshold=0.0
rec.recommender.class=usg
rec.alpha=0.1
rec.similarity.class=bcos
rec.recommender.similarities=user
rec.beta=0.1
rec.eta=0.05
rec.limit.userNum=101
rec.recommender.isranking=true
rec.recommender.ranking.topn=10
algorithmlist.txt · Last modified: 2018/08/05 01:27 by hrmvoc