摘要
为了实现高校图书馆借阅系统中的个性化推荐,本研究以图书的借阅持续时长、借阅总次数、续借次数作为兴趣度分量,利用协同过滤算法以及k近邻搜索算法解决借阅关系矩阵稀疏问题,构建基于兴趣度与类型因子的协同过滤推荐模型并设计了五层体系的书目推荐系统,实现了两大分区12个模块的借阅与推荐类功能。经过1 000名学生的实际借阅数据验证,结果表明当近邻个数取60以上且推荐书目为20时推荐效果最佳,为高校图书管理提供了智能化推荐手段。
In order to achieve the personalized recommendation in the university library borrowing system,this study takes book borrowing duration,lending total number,renew frequency as a scales,uses collaborative filtering algorithm and k nearest neighbor search algorithm to solve the problem that the lending relationship matrix is sparse,builds collaborative filtering recommendation model based on interest degree and types of factors.It designs the system with five layers,which realizes library and the recommended class function by the two partitions,12 modules.Through the actual borrowing data of 1000 students,the results show that the recommendation effect is the best when the number of neighboring books is more than 60 and the recommended books are 20,which provides an intelligent recommendation method for the library management of colleges and universities.
作者
赵峰涛
ZHAO Fengtao(Library,Xi’an Peihua University,Xi’an 710125,China)
出处
《微型电脑应用》
2022年第12期67-69,73,共4页
Microcomputer Applications
基金
西安培华学院2020年度校级课题(PHKT2024)。
关键词
高校图书馆
协同过滤算法
k近邻搜索算法
university library
collaborative filtering algorithm
k nearest neighbor search algorithm
