摘要
随着信息爆炸,个性化推荐技术成为提高信息检索效率和用户体验的关键。文章针对传统协同过滤推荐方法仅依赖用户静态评分导致推荐质量低的问题,提出基于用户评分多元关系分析的个性化推荐方法。该方法综合分析用户评分一致性、信任关系及用户行为,引入时间因素反映用户兴趣变化,建立用户间多元关系模型。实验结果表明,相较于其他推荐方法,文章方法在实际应用中能提供更精确的推荐服务,有效满足用户信息需求,实现了推荐质量的显著提升。
With the explosion of information,personalized recommendation technology has become key to improving information retrieval efficiency and user experience.This article addresses the issue of low recommendation quality caused by traditional collaborative filtering methods that rely solely on users'static ratings.It proposes a personalized recommendation method based on multi-dimensional analysis of user ratings.This method comprehensively analyzes user rating consistency,trust relationships,and user behavior,introduces temporal factors to reflect changes in user interests,and establishes a multi-dimensional relationship model among users.Experimental results show that compared to other recommendation methods,the approach presented in this article can provide more accurate recommendation services in practical applications,effectively meeting users'information needs and achieving a significant improvement in recommendation quality.
作者
张君
ZHANG Jun(Laboratory and Asset Management Division,Tianjin Medical University,Tianjin 300070)
出处
《江苏科技信息》
2025年第2期122-129,共8页
Jiangsu Science and Technology Information
基金
2021年度天津市教育科学规划课题,项目名称:基于用户画像的高校医学生信息素养教育研究,项目编号:HCE210338。
关键词
个性化推荐
用户评分
用户多元关系
用户行为量化
personalized recommendation
user ratings
user multi-relationship
quantifying user behavior
