《计算机应用研究》|Application Research of Computers

结合项目流行度加权的协同过滤推荐算法

Collaborative filtering recommendation algorithm combined with item popularity weighting

免费全文下载 (已被下载 次)  
获取PDF全文
作者 魏甜甜,陈莉,范婷婷,吴小华
机构 西北大学 信息科学与技术学院,西安 710127
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2020)03-008-0676-04
DOI 10.19734/j.issn.1001-3695.2018.08.0618
摘要 针对传统协同过滤算法中存在的流行度偏差问题,提出一种结合项目流行度加权的协同过滤推荐算法。在项目协同过滤算法的基础上,分析项目流行度和流行度差异对相似度的影响;设置流行度阈值,对大于该阈值的流行项目设计惩罚权重,降低其对项目间相似度的贡献。通过在MovieLens 1M和Epinion数据集上进行实验验证和对比,结果表明,所提算法的预测准确度和覆盖率均优于传统算法,有效提高了推荐的多样性和新颖性,一定程度上缓解了流行度偏差问题。
关键词 协同过滤; 相似性度量; 流行度偏差; 项目流行度
基金项目
本文URL http://www.arocmag.com/article/01-2020-03-008.html
英文标题 Collaborative filtering recommendation algorithm combined with item popularity weighting
作者英文名 Wei Tiantian, Chen Li, Fan Tingting, Wu Xiaohua
机构英文名 School of Information Science & Technology,Northwest University,Xi'an 710127,China
英文摘要 Aiming at the popularity bias problem in traditional collaborative filtering algorithms, this paper proposed a collaborative filtering recommendation algorithm combined with item popularity weighting. On the basis of the item collaborative filtering algorithm, it analyzed the influence of item popularity and popularity difference between items on similarity. This algorithm used the item popularity and popularity difference to set the penalty weight functions and adjusted the similarity between popularity items when the item popularity was greater than the threshold. The experiments on the MovieLens 1M and Epinion datasets show that the proposed algorithm has better prediction accuracy and coverage rate than traditional algorithms, which effectively improves the diversity and novelty of recommendations, and alleviates the popularity bias problem.
英文关键词 collaborative filtering; similarity measure; popularity bias; item popularity
参考文献 查看稿件参考文献
 
收稿日期 2018/8/14
修回日期 2018/10/9
页码 676-679
中图分类号 TP301.6
文献标志码 A