青岛大学学报（自然科学版）2025，Vol.38Issue(1)：37-42,50,7.DOI:10.3969/j.issn.1006-1037.2025.01.06
多维度自适应协同过滤的跨境电商产品信息推荐方法
Multidimensional Adaptive Collaborative Filtering Method for Recommendation of Cross Border E-commerce Product Information
摘要
Abstract
A multi-dimensional adaptive collaborative filtering method for cross-border e-commerce product information recommendation was proposed to address the shortcoming of the traditional recommendation method in terms of coverage and diversity.By compre-hensively considering the five dimensions of a user's browsing frequency,favorites,add-ons,purchase behavior,and comments,the product's rating by the user was calculated and a rating matrix was constructed.The adaptive k-means algorithm was adopted,a user was clustered,the similarity was calculated,the nearest neighbor was identified,and a product rating was predicted.Through this process,a recommendation list was generated and multi-dimensional adaptive collaborative filtering recommendation was achieved.The research results show that the multi-dimensional adaptive collaborative filtering recom-mendation method is better in terms of coverage,diversity,and user satisfaction,with the highest coverage reaching 0.88,the highest diversity reaching 0.9890,and the highest us-er satisfaction reaching 95.67%.关键词
多维度/自适应协同过滤/跨境电商产品/聚类/信息推荐方法Key words
multi-dimensional/adaptive collaborative filtering/cross border e-commerce products/clustering/information recommendation methods分类
信息技术与安全科学引用本文复制引用
钱远玲..多维度自适应协同过滤的跨境电商产品信息推荐方法[J].青岛大学学报（自然科学版）,2025,38(1):37-42,50,7.基金项目
2021年度福建省高校哲学社会科学研究项目(批准号:JAS21386)资助 （批准号:JAS21386）
2021年度泉州市社会科学规划项目(批准号:2021E01)资助. （批准号:2021E01）
