基于组合推荐技术的音乐推荐引擎-研究与实现

《基于组合推荐技术的音乐推荐引擎-研究与实现》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、万方数据广东工业大学硕士学位论文协同过滤算法。3．后台离线整合两种算法的推荐结果，在线推荐模块实时优化用户的推荐列表，从而实现个性化音乐推荐引擎系统。为了增强了用户的体验，』HjMongoDB加速数据的存取，用Spring搭建高效的后台系统，用Bootstrap和HTML5增强前台页面的视听效果，使我们的系统能应用在实际需求中。关键词：协同过滤；组合推荐；SimhashHadoop：MahoutⅡ万方数据Absn'actAbstractWhilepeoplegetmusicbyavarietyofways，mostofthemli

2、stenformtheInteract．Thebusymodemlifeandthevastnetworkresourcesmakealotofpeopletoobusytosearchtheirfavoritemusiccarefully,andalotofuserfavoritemusichasnochancetobeenjoyed．Howtodiscoveryuserfavorquicklyandhelpuserstofmdtheirfavoritemusic，whichisthejobofmusicrecommendati

3、onengine．RecommendationalgorithmistheCOreoftheengine，themeritsofthealgorithmdeterminesthequalityoftherecommendationresult．Theresearchofcontent-basedrecommendationalgorithmhasstartedinanearlytime．Thispaperuselabelstodescribemusicdata,makingitcanbeusedincontent—basedrec

4、ommendationalgorithmwhichismainlyclusteringalgorithm．thetraditionalTF·IDFalgorithmgeneratemusicdocumentvectorforclustering，notonlyhasalowefficiency，butabadrecommendationeffect．Therefore，thispaperputsforwardanewalgorithmforgeneratingvectors，whichuseSimhashalgorithmtocr

5、eatethefingerprintcharacteristicvaluefortheitemstOcluster．Thismethodgetsahighefficiencyandabetterclusteringeffectbytheexperiments．Inaddition，inthefieldofrecommendation，collaborativefilteringalgorithmisthemorewidelyappliedatpresent．Accordingtothecharacteristicsofthecol

6、laborativefilteringrecommendationalgorithmhaslargecalculationamount，thispapermainlytalksaboutimplementinguser—baseddistributedcollaborativefilteringalgorithmontheHadoopplatform，optimizingusermatrix，removinghotorcolditemsandsimplifyingthewholeprocess．Intheexperiments，c

7、omparedwithitem-baseddistributedcollaborativefilteringalgorithm，inthepremiseofsamedatasize，collaborativefilteringalgorithmhadafasterspeed．What’Smore，it’Squalitydidnotbecomeworse．Finally，thispaperbuiltamusicalrecommendationengineprototypesystemwithB／Sstructure，integrat

8、edtheresultofoffiinerecommendationalgorithmandaddedonlinerecommendationfunction，whichmeetsthereal．timeneedoftheusersandhelps