推荐算法

—— Recommender

欢马劈雪     最近更新时间:2020-08-04 05:37:59

438

Similarity

相似性矩阵用于构建数据集中user-user或者item-item之间的距离. 在使用命令行配置时, 示例配置如下:

rec.similarity.class=cos
rec.recommender.similarities=user

其中rec.similarity.class为指定相应的距离度量函数. rec.recommender.similarities为指定user-user之间的距离或者item-item之间的距离. 在Social类的推荐方法中, 还可以指定计算social-social之间的距离.

在Java程序中, 相应的示例程序如下

conf.set("rec.recommender.similarities","user");
RecommenderSimilarity similarity = new CosineSimilarity();
similarity.buildSimilarityMatrix(dataModel);

生成RecommenderSimilarity对象之后, 还需要调用buildSimilarityMatrix方法来进行相似度矩阵的计算.

下表为LibRec中已经实现的距离度量算法以及相应的简写.

Similarity shortname
BinaryCosineSimilarity bcos
CosineSimilarity cos
CPCSimilarity cpc
MSESimilarity msesim
MSDSimilarity msd
PCCSimilarity pcc
KRCCSimilarity krcc
DiceCoefficientSimilarity dice
JaccardSimilarity jaccard
ExJaccardSimilarity exjaccard

Algorithms

在使用配置项和命令行运行LibRec时, 执行的推荐算法由配置项rec.recommender.class指定.配置如下

rec.recommender.class=shortname #e.g. aobpr

不同算法的简写请参阅[Algorithm list.md](./Algorithm list)

在java实现中, 实例Configuration对象, DataModel对象, Similarity矩阵对象之后, 作为RecommenderContext的构造器参数生成RecommenderContext的对象. 此处可以直接实例相应的推荐算法类, 因此无需设置配置项rec.recommedner.class. 示例代码如下:

RecommenderContext context = new RecommenderContext(conf, dataModel, similarity);

conf.set("rec.neighbors.knn.number","50");
conf.set("rec.recommender.isranking=false");

Recommender recommender = new UserKNNRecommender();
recommender.recommend(context);

推荐算法根据不同的计算方式具有各自相应的配置项. 目前LibRec中使用的推荐算法分别有基于矩阵分解算法, 基于因子分解机算法, 概率图模型, 基于张量的算法等. 下面依次给出基于不同算法的推荐算法配置项. 一般来说, 以矩阵分解为例, 继承矩阵分解接口的推荐算法在进行推荐计算时除去配置矩阵分解的配置项之外, 还需要配置其他配置项. 如BPMF算法中,还需要配置rec.recommender.user.mu等配置项.

实现基于以上算法的推荐系统只需继承相应的抽象类即可. 不同算法的配置项列举在下一栏, 目前已经实现的所有推荐算法配置项列举在Algorithm List中.

AbstractRecommender

# if ranking
rec.recommender.isranking=true
rec.recommender.ranking.topn=10

直接实现AbstractRecommender抽象类的算法有

directory path short name algorithm
baseline constantguess ConstantGuessRecommender
baseline globalaverage GlobalAverageRecommender
baseline itemaverage ItemAverageRecommender
baseline mostpopular MostPopularRecommender
baseline randomguess RandomGuessRecommender
baseline useraverage UserAverageRecommender
cf.rating rbm RBMRecommender
cf itemknn ItemKNNRecommender
cf userknn UserKNNRecommender
cf.ranking slim SLIMRecommender
ext associationrule AssociationRuleRecommender
ext external ExternalRecommender
ext personalitydiagnosis PersonalityDiagnosisRecommender
ext slopeone SlopeOneRecommender
hybrid hybrid HybridRecommender

Probabilistic Graphical Recommender

rec.iterator.maximum=1000
rec.pgm.burn-in=100
rec.pgm.samplelag=100

直接继承自Probabilistic Graphical Recommender的算法有

directory path short name algorithm
baseline itemcluster ItemClusterRecommender
baseline usercluster UserClusterRecommender
cf bhfree BHFreeRecommender
cf bucm BUCMRecommender
cf.ranking aspectmodelranking AspectModelRecommender
cf.ranking itembigram ItemBigramRecommender
cf.ranking lda LDARecommender
cf.ranking plsa PLSARecommender
cf.rating aspectmodelrating AspectModelRecommender
cf.rating gplsa GPLSARecommender
cf.rating ldcc LDCCRecommender
cf.rating urp URPRecommender

Matrix Factorization Recommender

rec.iterator.maximum=100
rec.iterator.learningrate=0.01
rec.iterator.learningrate.maximum=1000
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learningrate.bolddriver=false
rec.learningrate.decay=1.0

直接或间接继承Matrix Factorization Recommender的推荐算法有

directory path short name algorithm
cf.rating asvdpp ASVDPlusPlusRecommender
cf.rating svdpp SVDPlusPlusRecommender
content efm EFMRecommender
content hft HFTRecommender
context.rating timesvd TimeSVDRecommender
cf.ranking aobpr AoBPRRecommender
cf.ranking bpr BPRRecommender
cf.ranking climf CLIMFRecommender
cf.ranking eals EALSRecommender
cf.ranking fismauc FISMaucRecommender
cf.ranking fismrmse FISMrmseRecommender
cf.ranking gbpr GBPRRecommender
cf.ranking listwisemf ListwiseMFRecommender
cf.ranking rankals RankALSRecommender
cf.ranking ranksgd RankSGDRecommender
cf.ranking wbpr WBPRRecommender
cf.ranking wrmf WRMFRecommender
cf.rating biasedmf BiasedMFRecommender
cf.rating bnpoissmf BNPoissMFRecommender
cf.rating bpmf BPMFRecommender
cf.rating bpoissmf BPoissMFRecommender
cf.rating llorma LLORMARecommender
cf.rating mfals MFALSRecommender
cf.rating nmf NMFRecommender
cf.rating pmf PMFRecommender
cf.rating rfrec RFRecRecommender

Factorization Machine Recommender


rec.recommender.maxrate=12.0
rec.recommender.minrate=0.0

rec.factor.number=10

rec.fm.regw0=0.01
reg.fm.regW=0.01
reg.fm.regF=10

直接继承自Factorization Machine Recommender的算法有

directory path short name algorithm
cf.rating fmals FMALSRecommender
cf.rating fmsgd FMSGDRecommender

Social recommender

rec.iterator.maximum=100
rec.iterator.learningrate=0.01
rec.iterator.learningrate.maximum=1000
rec.user.regularization=0.01
rec.item.regularization=0.01
rec.factor.number=10
rec.learningrate.bolddriver=false
rec.learningrate.decay=1.0

rec.social.regularization=0.01

直接继承自Social recommender的算法有

展开阅读全文