作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2013, Vol. 39 ›› Issue (1): 58-62. doi: 10.3969/j.issn.1000-3428.2013.01.012

• 软件技术与数据库 • 上一篇    下一篇

基于项目和信任的协同过滤推荐算法

朱丽中,徐秀娟,刘 宇   

  1. (大连理工大学软件学院,辽宁 大连 116620)
  • 收稿日期:2012-02-15 修回日期:2012-05-21 出版日期:2013-01-15 发布日期:2013-01-13
  • 作者简介:朱丽中(1988-),女,硕士研究生,主研方向:数据挖掘,推荐算法;徐秀娟,讲师、博士;刘 宇,副教授、 博士后
  • 基金资助:
    国家自然科学基金资助项目(60803074, 60673024);大连理工大学引进人才科研启动基金资助项目

Collaborative Filtering Recommendation Algorithm Based on Item and Trust

ZHU Li-zhong, XU Xiu-juan, LIU Yu   

  1. (School of Software Technology, Dalian University of Technology, Dalian 116620, China)
  • Received:2012-02-15 Revised:2012-05-21 Online:2013-01-15 Published:2013-01-13

摘要: 为解决冷启动用户的推荐问题,对TrustWalker算法在相似度计算、可能性项目选择和预测评分等方面进行改进,提出一种基于项目和信任的协同过滤推荐算法CoTrustWalker。采用云模型相似度方法计算项目间的相似度,通过选择最相似的若干个项目的聚合结果作为随机游走的返回结果,从而提高推荐结果的稳定性。实验结果表明,CoTrustWalker算法在小规模数据集上与TrustWalker算法相比,其推荐质量和推荐速度均有较大提高。

关键词: 推荐系统, 协同过滤, 基于信任网络推荐, 冷启动推荐, 混合推荐, 云模型

Abstract: To solve recommendation problem for cold-start users, this paper proposes a collaborative filtering algorithm CoTrustWalker that combines item-based collaborative filtering and trust-based recommendation, which improves TrustWalker algorithm in similarity calculation, the possibility of item selection and prediction score. CoTrustWalker algorithm adopts the cloud model similarity method to compute the item similarities, so as to enhance the stability of recommendation results by selecting the most similar polymerization results of some items. Experimental results show that CoTrustWalker algorithm improves the quality of recommendation and the speed compared with TrustWalker algorithm.

Key words: recommendation system, collaborative filtering, trust-based network recommendation, cold start recommendation, hybrid recommendation, cloud model

中图分类号: