期刊文献+

农业电子商务中基于用户兴趣变化的协同过滤推荐技术研究 预览

Research on Collaborative Filtering Recommendation Technology Based on Users' Interest Change in Agricultural E-commerce
在线阅读 下载PDF
收藏 分享 导出
摘要 由于农业本身季节性强的特点,推荐系统的实时性对农业电子商务网站至关重要。然而传统的协同过滤推荐技术不可避免地存在时效性差的问题。针对这一问题,本文分别从用户对商品的浏览时间及浏览频率两个方面追踪用户兴趣的变化,提出了浏览时间权重和浏览率权重,进而设计用户兴趣权重。将反映用户兴趣的权重值引入传统的协同过滤推荐技术中,得到新的加权计算用户相似度的计算方法,使推荐结果可以随着用户兴趣的迁移而改变。实验结果证明,该方法大大提高了推荐结果的准确性。 Due to seasonal characteristics of agriculture itself,real-time performance of recommendation system is very important for agricultural e-commerce website.However,the traditional collaborative filtering recommendation technology inevitably had problems of poor timeliness.In order to solve this problem,the change of users' interest was tracked separately from two aspects including the user's browsing time and browsing rate.The weight of browsing time and browsing rate were proposed,and then the weight of user interest was designed.The weight value of user's interest was introduced into the traditional collaborative filtering recommendation technology,and a new method for calculating users similarity with the weight was obtained.And then the recommendation results can be changed with the migration of user's interest.Experimental results showed that this method can greatly improve the accuracy of results.
作者 程芳 CHENG Fang (College of Information Science and Technology, Agricultural University of Hebei, Hebei Baoding 071000)
出处 《农业网络信息》 2016年第5期41-44,47共5页 Agriculture Network Information
基金 保定市科学技术研究与发展指导计划项目“基于协同过滤的农业信息推荐系统的研究与开发”(编号:14ZN019)
关键词 农业电子商务 推荐系统 协同过滤 兴趣变化 实时性 agricultural e-commerce recommending system collaborative filtering interest change real-time performance
作者简介 程芳(1980-),女,硕士,讲师,研究方向:数值计算与计算机应用。
  • 相关文献

参考文献18

二级参考文献255

共引文献654

投稿分析

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部 意见反馈
新型冠状病毒肺炎防控与诊疗专栏