基于大数据的高校图书馆个性化推荐书目生成研究

Research on Personalized Recommendation Bibliography Generation of University Library Based on Big Data

通过数据挖掘技术可以发现在校读者的借阅习惯以及使用图书资源的模式,进而评估读者对馆藏资源和馆藏服务的利用情况,针对读者的借阅规律,图书馆可以提供个性化的信息推送服务,有效提高资源利用率和服务水平。以辽宁师范大学文、史、法及心理学院读者的借阅记录为样本数据,采用大数据处理软件Weka进行数据离散化转换,并加载分析,根据频繁项集合算法的挖掘关联规则,预测相关书籍的借阅概率,生成推荐书目,向读者进行个性化推荐。经过大数据分析发现,读者借阅同种图书的关联度占总关联规则的比率较大,说明大部分读者在一次特定的借阅中,往往只会借阅某一类别或者高度相关的图书。将上述结果提供给相关学科馆员,能为读者提供更有针对性和目的性的书目,并加以个性化信息推送服务,提高图书馆的学科服务质量。

Through the data mining technology,we can find reader＇s borrowing habits in the school and the model of books resources utilization,and then assess collection of resources and collection of services for the readers. For readers＇ borrowing pattern,the library can provide personalized information pushing service,which can effectively improve resource utilization and service levels. This paper uses the big data processing software Weka to carry out data discretization and load analysis,and according to the mining association rule of frequent item collection algorithm,the probability of borrowing related books are predicted according to the borrowing records of readers from history,law and psychology department of Liaoning Normal University,and then generates recommended bibliography. Through the analysis of big data,it is found that the ratio of readers＇ borrowing of the same kind of books is larger than that of the total association rules,which indicates that most readers will only borrow certain categories or highly related books in a specific borrowing. The above results can be provided to the relevant subject librarians,to provide readers with more targeted and purpose books for information pushing service,and improve the quality of library services.