摘要
从功能目标、设计方案、模型设计、建模过程、评估方法对高校图书馆阅读书目智慧推荐系统进行分析,以最小成本实现阅读推广服务效益的最大化。推荐系统模型由图书馆自动化系统平台、教育机构信息系统、电子图书数据库、DDA购书平台、推荐模块组成。以Hadoop技术收集分析3个不同源的数据,使用MapReduce搭建模型框架,通过Mahout算法实现基于共现矩阵的图书相似度推荐,利用皮尔森相似度计算公式推荐图书。读者可以收到推荐系统给予的3种推荐:馆内预约借阅、全文电子书阅读与下载、图书购买。在修改推荐算法、开发微信小程序、整合图书馆物联网数据、更改DDA交互模式方面可以对推荐系统进行优化。
From the functional objectives,design schemes,model design,modeling process,and evaluation methods,the intelligent recommendation system for reading bibliographies of college libraries is analyzed to maximize the benefits of reading promotion services at the minimum cost.The recommendation system model is composed of library automation system platform,educational institution information system,electronic book database,DDA book purchase platform,and recommendation module.It uses Hadoop technology to collect and analyze data from three different sources,uses MapReduce to build a model framework,implements book similarity recommendation based on co-occurrence matrix through Mahout algorithm,and recommends books by using Pearson similarity calculation formula.Readers can receive three types of recommendations from the recommendation system:in-library reservation,full text e-book reading and downloading,and book purchase.The recommendation system can be optimized in terms of modifying the recommendation algorithm,developing WeChat applets,integrating library Internet of Things data,and changing the DDA interaction mode.
出处
《图书馆学研究》
CSSCI
北大核心
2020年第23期91-100,F0003,共11页
Research on Library Science
基金
2020年度中国图书馆学会阅读推广课题“基于智慧服务的图书馆阅读推广服务研究”(项目编号:YD2020B42)的研究成果之一。
关键词
高校图书馆
智慧推荐
推荐系统
阅读推广
college library
intelligent recommendation
recommendation system
reading promotion