摘要
如何为用户提供感兴趣的个性化推荐图书商品向来都是智慧图书馆最核心的难题之一。因此,利用优质的推荐算法构建推荐系统就变得尤为重要。基于多维关系和用户聚类两个方面构建推荐算法,在充分考虑用户之间相关关系和图书商品之间相关关系的基础上,不断更新“用户-图书商品”二维矩阵,使其数值更加合理和真实。随后使用k均值聚类方法聚拢高相关性用户。最后在类内选取目标用户实现个性化图书商品推荐。实验结果表明,该推荐算法能取得较高的评价指标F1值,即算法更加优质和有效。
How to provide users with personalized recommended books has always been one of the most concerned problems of intelligent library.Therefore,it is very important to use high quality recommendation algorithm to build recommendation system.The recommendation algorithm is constructed from the two aspects of multidimensional relationship and user clustering.On the basis of fully considering the correlation between users and the relationship between books and commodities,the two‑dimensional matrix of“user‑book commodities”is constantly updated to make its value more reasonable and real.Then,the K‑means clustering method was used to gather high correlation users.Finally,select target users within the class to realize personalized book product recommendation.The experimental results show that the recommendation algorithm can obtain a higher value of the evaluation index F1,and the algorithm is more high‑quality and effective.
作者
闫俊辉
Yan Junhui(School of Mathematics and Information Technology,Yuncheng University,Yuncheng 044000,China)
出处
《现代计算机》
2023年第14期62-65,73,共5页
Modern Computer
关键词
智慧图书馆
多维关系
用户聚类
个性化
推荐系统
intelligent library
multidimensional relationship
user clustering
personalization
recommendation system