基于情境感知的电商平台推荐系统框架研究
The Framework of E-commerce Platform Recommendation System Based on Context-awareness
[目的/意义] 随着互联网行业的蓬勃发展,电商平台的用户与商品数量均呈现井喷增长。这一趋势在促进互联网经济发展的同时也为用户带来了“信息过载”的问题。电商平台如何从用户的历史行为日志中获取信息,并根据用户当前情境推荐商品,不仅是重要的理论与技术课题,同时也蕴藏着巨大的商业价值。[方法/过程]文章结合情境感知技术,提出了一种五层推荐系统框架,自底向上分为:感知层、过滤层、排序层、规则层和应用层。同时,利用淘宝 APP 的购物数据进行了模拟推荐实验,比较了各层次不同设计方式的推荐效果。[结果/结论] 研究发现,融入了情境感知技术的推荐系统框架可以显著提高推荐的精确率, 但是会削弱召回率。基于此,本研究为电商平台提出了部分建议。
[Purpose/significance] With the rapid development of internet industry, the numbers of users and commodities are gaining explosive growth. This trend not only boosts the development of internet economy, but also brings the “information overload” problem. It makes mining latent information from logs on basis of context awareness not only an important research topic, but also a great commercial value to e-commerce platforms. [Method/process] This study proposed a new context-recommendation hierarchical framework, which has 5 bottom-up layers. We have compared the performances of each procedure of this framework. [Result/conclusion] The result shows the context-awareness can improve the precision of recommendation significantly at the expense of lowering the recall. Furthermore, we provide some reasonable advice for e-commerce platforms.
张孜铭、程秀峰
科学交流与知识传播
推荐系统情境感知电商平台用户行为系统分析
recommend system context awareness e-commerce platform user behavior system analysis
中国博士后科学基金第 65 批面上项目 基于情境感知的资源发现系统功能优化策略研究( 2019M650802 ) 国家自然科学基金项目 基于情境感知的智慧图书馆阅读与交流服务实现路径研究( 71974069 )
张孜铭,程秀峰.基于情境感知的电商平台推荐系统框架研究[EB/OL].(2022-07-04)[2024-12-22].https://sinoxiv.napstic.cn/article/3444813.点此复制
评论