Please wait a minute...
FITEE
Front. Inform. Technol. Electron. Eng.  2016, Vol. 17 Issue (6): 489-500    DOI: 10.1631/FITEE.1500313
    
基于兴趣转移模型的协同过滤算法
Bin Ju, Yun-tao Qian, Min-chao Ye
Institute of Artificial Intelligence, College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China; Health Information Center of Zhejiang Province, Hangzhou 310006, China
Preference transfer model in collaborative filtering for implicit data
Bin Ju, Yun-tao Qian, Min-chao Ye
Institute of Artificial Intelligence, College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China; Health Information Center of Zhejiang Province, Hangzhou 310006, China
 全文: PDF 
摘要： 目的：由于单一域中的用户-物品关系数据的极度稀疏性，很容易遇到预测的“天花板”。为提高隐性关系数据的预测效果，把来自浏览域的用户-物品频次矩阵与来自购物域的用户-物品频次矩阵进行整合，能够提早预测到用户兴趣的时间发生点，从而提高推荐系统的预测精度。
创新点：采用多任务非负矩阵分解框架。不同于前人基于高斯分布先验的矩阵分解模型，本文基于泊松分布先验，提出一种基于用户兴趣因子转移的跨域的非负矩阵分解模型。
方法：首先，“先看后买”是人类购物的一般行为。来自浏览行为背后的用户兴趣一定早于购物行为背后的用户兴趣发生，因此通过概率图模型对矩阵进行共享物品因子的分解（图2）。然后，针对用户-物品关系矩阵中的数据是频次数据的特点，提出了一个基于泊松分布先验的多任务非负矩阵分解算法（算法1）。算法第一阶段，把用户-物品浏览矩阵和用户-物品购买矩阵一起分解为用户浏览兴趣因子矩阵和用户购买兴趣因子矩阵。第二阶段，把两个兴趣因子矩阵按列的最大值置1，其余为0，然后做两个因子矩阵的点积，生成所谓的兴趣转移字典。第三阶段，根据兴趣转移字典重构下一阶段用户-物品的购买矩阵，预测未来用户可能会购买何种物品（图3）。
结论：基于相同用户、相同物品的不同业务场景信息（如浏览行为数据和购买行为数据）分析用户潜在兴趣的概率产生模式，能够大幅度提升预测用户未来购买何种物品的效果。
关键词： 推荐系统;  协同过滤;  兴趣转移模型;  跨域;  隐性数据    
Abstract: Generally, predicting whether an item will be liked or disliked by active users, and how much an item will be liked, is a main task of collaborative filtering systems or recommender systems. Recently, predicting most likely bought items for a target user, which is a subproblem of the rank problem of collaborative filtering, became an important task in collaborative filtering. Traditionally, the prediction uses the user item co-occurrence data based on users’ buying behaviors. However, it is challenging to achieve good prediction performance using traditional methods based on single domain information due to the extreme sparsity of the buying matrix. In this paper, we propose a novel method called the preference transfer model for effective cross-domain collaborative filtering. Based on the preference transfer model, a common basis item-factor matrix and different user-factor matrices are factorized. Each user-factor matrix can be viewed as user preference in terms of browsing behavior or buying behavior. Then, two factor-user matrices can be used to construct a so-called ‘preference dictionary’ that can discover in advance the consistent preference of users, from their browsing behaviors to their buying behaviors. Experimental results demonstrate that the proposed preference transfer model outperforms the other methods on the Alibaba Tmall data set provided by the Alibaba Group.
Key words: Recommender systems    Collaborative filtering    Preference transfer model    Cross domain    Implicit data
收稿日期: 2015-09-25 出版日期: 2016-06-06
CLC:  TP391  
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章　　
Bin Ju
Yun-tao Qian
Min-chao Ye

引用本文:

Bin Ju, Yun-tao Qian, Min-chao Ye. Preference transfer model in collaborative filtering for implicit data. Front. Inform. Technol. Electron. Eng., 2016, 17(6): 489-500.

链接本文:

http://www.zjujournals.com/xueshu/fitee/CN/10.1631/FITEE.1500313    或    http://www.zjujournals.com/xueshu/fitee/CN/Y2016/V17/I6/489

[1] Zhen-ming Yuan, Chi Huang, Xiao-yan Sun, Xing-xing Li, Dong-rong Xu. 基于社会标签和时间兴趣演变模型的微博推荐算法[J]. Front. Inform. Technol. Electron. Eng., 2015, 16(7): 532-540.
[2] Ming Yang, Ying-ming Li, Zhongfei (Mark) Zhang. 基于主题回归和关联矩阵分解的科技文献推荐[J]. Front. Inform. Technol. Electron. Eng., 2014, 15(11): 984-998.