图书馆大数据知识服务情境化推荐系统研究  

Study on Situational Recommendation System of Library Big Data Knowledge Service

作　　者：刘海鸥[1,2] 陈晶[1] 孙晶晶[1] 张亚明[1,2] Liu Hai-ou;Chen Jing;Sun Jing-jing;Zhang Ya-ming

机构地区：[1]燕山大学互联网+与产业发展研究中心 [2]燕山大学经济管理学院

出　　处：《图书馆理论与实践》2018年第8期98-103,共6页Library Theory and Practice

基　　金：国家社科基金"基于大数据深度融合的移动图书馆用户画像情境化推荐模型研究"(项目编号:18BTQ033)经费资助研究成果之一

摘　　要：大数据环境下,推荐系统项目评分的稀疏性问题愈加突出,如何从大数据中挖掘出与用户情境精准匹配的知识服务资源并及时推荐给相应用户,成为图书馆大数据知识服务亟待解决的问题。文章在系统回顾相关研究进展的基础上,提出图书馆大数据知识服务的情境化推荐系统框架结构,并通过融合图书馆用户的情境信息进行大数据知识服务的协同过滤推荐,由此缓解数据稀疏性导致的推荐性能下降问题;同时采用Map Reduce的并行处理方式,以提高大数据的并行挖掘性能。In the Big Data environment,the sparseness of recommendation system is becoming more and more prominent.How to excavate the knowledge service resources that match exactly with user scenarios and recommend them to users timely become the urgent problem to be solved in library Big Data knowledge service.Based on a review of related research,it establishes a recommendation system framework of Big Data knowledge service.Meanwhile,the collaborative filtering recommendation of Big Data knowledge service is carried out by integrating the situation information of library users,so as to alleviate the recommendation performance degradation caused by data sparseness.The parallel processing method of MapReduce is adopted to improve the parallel mining performance of Big Data.The research results provide some new perspectives for improving the service level of Big Data knowledge in library.

关 键 词：情境化推荐 大数据知识服务 图书馆 协同过滤 

分 类 号：G250.73[文化科学—图书馆学] G252