《推荐系统(大学英语课堂报告)》由会员分享，可在线阅读，更多相关《推荐系统(大学英语课堂报告)（35页珍藏版）》请在金锄头文库上搜索。
1、推荐系统By xxxVideoBlogMusicTravelingSearchingNewsOnline ShoppingMapPictureEventChattingSMSEmailRecommender System石家庄铁道大学 电子商务 专业英语报告Recommender System is a specific type of information filtering system that attempts to recommend information items that are likely to be of interest to the users.Recommend
2、er System石家庄铁道大学 电子商务 专业英语报告l Must fulfil users demands l The data generated is beneficial for development l Must fulfil owners demandsUserRecommender SystemOwnerDataDataRec.Recommender System石家庄铁道大学 电子商务 专业英语报告l Collaborative filtering recommendation 协同过滤推荐 l Content-based recommendation 基于内容的推荐 l
3、Knowledge-based recommendation 基于知识的推荐 l Utility-based recommendation 基于效用的推荐 l Demographic-based Recommendation 基于用户统计信息的推荐 l Association rule-based recommendation 基于关联规则的推荐 l Hybrid recommendation 混合推荐Recommender System石家庄铁道大学 电子商务 专业英语报告Collaborative filtering recommendation协同过滤推荐lUser-based CF r
4、ecommendation 基于用户的协同过滤推荐 lItem-based CF recommendation 基于物品的协同过滤推荐Recommender System石家庄铁道大学 电子商务 专业英语报告User-based CFl User-item matrix l Find similar users l Compute the similarity of two users items l Recommend similar users choices l Asking friends for a recommendation l User like items of user w
5、ho share similar interestsItem-based CFl User-item matrix l Find similar items l Compute the similarity of two items for all users l Recommend items which have high- similarity l People who buy x also buy y l People like similar stuff which they like beforeRecommender System石家庄铁道大学 电子商务 专业英语报告Conten
6、t-based recommendation基于内容的推荐lUser evaluate the objects lThe system learn users interests lPredict if a new object match users interests lThen recommend to userRecommender System石家庄铁道大学 电子商务 专业英语报告Demographic-based recommendation基于用户统计信息的推荐lAccording to the users personal attributes divide them into
7、 classes lGive different recommendations to different classes of users lIt doesnt require a history of user dataRecommender System石家庄铁道大学 电子商务 专业英语报告CEO of the AmazonRecommender System石家庄铁道大学 电子商务 专业英语报告Function one：transforming the Web site visitors into buyers . Site visitors always dont have desi
8、re to buy. But recommender system can recommend goods that the user may be interesting in, which can led to purchase process.Recommender System石家庄铁道大学 电子商务 专业英语报告Function two: to improve the cross-selling of the e-commerce site. recommender system to provide users with other valuable goods informati
9、on. So users can buy commodity that they are really in need but dont think of. Recommender System石家庄铁道大学 电子商务 专业英语报告Function three: to enhance customer loyalty for e-commerce sites. Recommender system analyze the users buying habits, and according to the user needs to provide users with valuable goo
10、ds. Therefore, recommender system is not only able to provide users with personalized service, but also to establish long-term stable relationship with the user.Recommender System石家庄铁道大学 电子商务 专业英语报告Problems :lLack of DatalChanging DatalChanging User PreferencelUnpredictable ItemslThe Stuff is Comple
11、xHome PageSearch Result PageGoods Details PageShopping Cart PageOrder Completed PageRecommender System石家庄铁道大学 电子商务 专业英语报告More informations offered : lTime (时间) lLocation (地点) lOnline/Offline (活动状态) lWeather (天气) lDevice type (设备类型) lContects (电话薄) lBinded account (绑定的账号)Recommender System石家庄铁道大学 电子商
12、务 专业英语报告Anywhere, AnytimeRecommender System石家庄铁道大学 电子商务 专业英语报告More easily to be accepted lAccuracy (精确的) lUseful (有用的) lRealtime (实时的) lCustomized (定制的)Reference1 (百分点)2佚名. A study of recommender System. 华中科技大 学 20113 gary. 推荐系统五大问题. 4 吴兵，叶春明. 基于效用的个性化推荐方法. J 计算机 工程 2012-2 第38卷4期5 孟祥武，胡勋，王立才，张玉洁. 移动推荐系统及其应 用研究. J 软件学报 2012-8
