贝叶斯模型大数据分析的软件实现——以河北科技大学图书馆为例

Implementation of Library Big Data Analysis Using a Bayes Network Toolbox——A Case Study of Hebei University of Science and Technology Library

文章利用贝叶斯模型整合图书馆的读者信息和借阅记录,用以预测读者的借阅行为,为读者提供图书借阅的参考,并对图书购买和使用提出建议。文章用MATLAB的贝叶斯工具箱FULLBNT1.0.7训练出用于借阅行为分析的贝叶斯模型bnet,开发软件Borrowbehavior实现了这种行为分析的功能,也对bnet中的概率矩阵进行知识挖掘。Borrowbehavior可根据读者信息给出读者借阅22类图书的可能性,bnet的5个节点之间的关联知识可给出关于图书馆建设的一些优化建议。

Libraries could use the Bayes model to integrate readers’information and their borrowing records,and predict their future borrowing behaviors,so as to provide better service for readers and in turn improve book acquisition.The MATLAB toolbox FULLBNT1.0.7 is used to create a Bayesian model bnet,and then a software named Borrowbehavior is developed to analyze readers’borrowing behaviors.Besides,a knowledge mining is carried on in the probability matrix in bnet.Borrowbehavior could predict 22 types of reader’s future borrowing possibilities based on their information and borrowing records,and bnet could provide some suggestions on library development.Such are helpful for libraries to improve their information service.