您的位置:首页  > 论文页面

基于扩散的推荐算法的可预测性

发表时间:2021-12-31  浏览量:591  下载量:63
全部作者: 王玫申,张鹏,薛乐洋
作者单位: 北京邮电大学理学院;北京师范大学复杂系统国际科学中心
摘 要: 推荐系统能够帮助解决信息过载问题,并可以根据用户的历史记录预测其兴趣进行推荐。推荐系统的核心是推荐算法,而提高最大推荐精度对进一步提升推荐算法精度至关重要。针对这一问题,提出了一个添加虚拟边的方法,通过增加推荐算法资源扩散的宽度使更多的商品可以获得资源,提高测试集内商品出现在推荐列表中的可能性,从而提高基于扩散的推荐算法的最大推荐精度。实验结果表明,本文方法提升了基于扩散的推荐算法的可预测性,同时也提升了推荐的准确性。
关 键 词: 应用数学;推荐系统;扩散算法;虚拟边;可预测性
Title: Predictability of diffusion-based recommendation algorithm
Author: WANG Meishen, ZHANG Peng, XUE Leyang
Organization: School of Science, Beijing University of Posts and Telecommunications; International Academic Center of Complex Systems, Beijing Normal University
Abstract: Recommendation system can help solve the problem of information overload, and can predict the interests of users according to their historical records for recommendation. The core of the recommendation system is recommendation algorithm, improving the maximum recommendation accuracy is very important for further improving the accuracy of the recommendation algorithm. In order to solve this problem, a method of adding virtual edges was proposed in this paper. By increasing the width of resource diffusion, more products could obtain resources, and the possibility of products in test set appearing in recommendation list was improved, so as to improve the maximum recommendation accuracy of the diffusion-based recommendation algorithm. Experimental results show that this method improves the predictability of the diffusion-based recommendation algorithm and the accuracy of recommendation.
Key words: applied mathematics; recommendation system; diffusion algorithms; virtual edge; predictability
发表期数: 2021年12月第4期
引用格式: 王玫申,张鹏,薛乐洋. 基于扩散的推荐算法的可预测性[J]. 中国科技论文在线精品论文,2021,14(4):462-467.
 
9 评论数 0
暂无评论
友情链接