计算机工程与应用 ›› 2021, Vol. 57 ›› Issue (18): 163-171.DOI: 10.3778/j.issn.1002-8331.2005-0363

• 模式识别与人工智能 • 上一篇    下一篇

融合情境语义推理及社会网络的团购推荐研究

李金海,何有世,张鹏   

  1. 1.泰州学院 计算机科学与技术学院,江苏 泰州 225300
    2.江苏大学 管理学院,江苏 镇江 212013
    3.泰州学院 经济与管理学院,江苏 泰州 225300
  • 出版日期:2021-09-15 发布日期:2021-09-13

Research of Group Recommendation Based on Contextual Semantics Reasoning and Social Network

LI Jinhai, HE Youshi, ZHANG Peng   

  1. 1.School of Computer Science and Technology, Taizhou University, Taizhou, Jiangsu 225300, China
    2.School of Management, Jiangsu University, Zhenjiang, Jiangsu 212013, China
    3.School of Economics and Management, Taizhou University, Taizhou, Jiangsu 225300, China
  • Online:2021-09-15 Published:2021-09-13

摘要:

随着团购以及移动终端的发展越演越烈,传统的推荐机制在团购推荐问题上逐渐呈现出弊端。针对团购推荐的情境敏感性,设计了基于情境语义推理的偏好分析模块,综合考虑了消费者长期演化形成的长期偏好以及基于情境形成的短期偏好;并将团购网络中的每个成员设定为复杂网络中的一个节点,将成员间的相互关系表示为节点的链接,构建团购用户的复杂网络,并设计了基于复杂网络的社会影响分析模块,以进行团购社区的划分,以及评测社区中成员之间的影响。基于此,完成了团购平台中两类产品的团购推荐研究。仿真实验表明,融合情境语义推理及社会网络的团购推荐机制具有良好的有效性及用户反馈满意度。

关键词: 情境推理, 复杂网络, 用户偏好, 团购推荐, 社会影响

Abstract:

As the development of group purchase and mobile terminal is escalating, the traditional recommendation mechanism on the problems of group recommendation gradually appeares disadvantages. In view of the context sensitivity of the group recommendation, this paper designs a preference analysis module based on the contextual semantics reasoning. It considers the long-term preference of consumers which form for a long time and short-term preference which form based on context. Each member of the group network is set to a node in complex network and the relationship between members are represented as the link of nodes. Then complex network of group users is built. This paper designs a social impact analysis module based on complex network. It uses for the division of community on group purchase and evaluation influence among members of the community. Based on this, the group recommendation study of two kind products on group platform is completed. Simulation results show that group recommendation mechanism based on contextual semantics reasoning and social network has a good validity and user feedback satisfaction.

Key words: context reasoning, complex network, user preference, group recommendation, social influence