导　　师： 陈琼;邢豫

学科专业： 0812

授予学位： 硕士

作　　者： ;

机构地区： 华南理工大学

摘　　要： 随着数据库技术及网络通信技术的高速发展,当前大多数高校图书馆都实现了从传统人工服务模式向计算机自动化处理模式的转变,其中图书的借还和编目流程几乎都通过数据库技术进行管理。经过多年的积累,图书馆的图书管理系统里已经存有大量信息,若能发掘隐藏在其中的某些有价值信息或者某些关联规律,那么将对图书馆的管理和服务起到很大的指导意义。 数据挖掘技术在图书馆管理系统的应用有着十分重要的意义,它能很好的从读者的借阅历史记录中发现师生们的阅读兴趣,发现借阅规律,从而为读者推荐相关图书,达到加强图书馆资源整合与利用,调整藏书和图书的摆放格局以方便师生借阅的目的,充分地有效地应用这种技术,对于及时了解读者信息需求,改进图书馆服务质量,促进图书管理等都具有非常强的指导意义。 关联规则和协同过滤是数据挖掘研究领域中常用的两种重要技术,关联规则挖掘能发现大量数据中项集之间有趣的关联或相关联系,协同过滤挖掘则利用具有相似兴趣的用户或具有相似属性的物品来实现相关推荐功能。本文详细介绍了如何运用关联规则和协同过滤挖掘算法对图书馆图书借阅记录进行信息挖掘,并以笔者所在图书馆的真实借阅记录数据为基础,运用基于用户和图书的协同过滤技术设计出图书推荐系统,挖掘出了读者之间借阅爱好的关联性以及图书之间的关联性,得出读者了借阅规律,实现为读者提供个性化图书推荐功能,最后通过实验评估了推荐系统的有效性和预测性,对于指导图书馆在图书推荐工作具有一定的参考价值。 With the development of data bases and internet communications technology, most libraries in higher education institutions have transformed from the manual service mode to the computer-assisted service, especially in the field of book borrowing and returning and book cataloging. With the long period of application, the library bibliographic retrieval system has accumulated a large amount of information; it would improve the management and service if we can find out the collaborative rules behind that information. Data mining technology is vital in the application of library bibliographic retrieval system, it can keep the track of the readers'interests and trace the book borrowing rules, hence to recommend books to readers with relevant topics and make integrations of all the library resources. Using data mining technology effectively, we can better understand the needs of readers and improve the service, as well as upgrade the book management. Association rules and collaborative filtering are the two vital technologies in the field of data mining research. Association rules is to find out the relations among items in a large amount of information, collaborative rules can realize recommendation function by collaborating users with similar interests or items with similar attributes. This paper thoroughly explains how to proceed the data mining with the record of book borrowing and returning by using the association rules and collaborative filtering rules. Based on the authentic record of book borrowing and retuning, this paper manages to design a book recommendation system with reference to the user-based collaborative filtering technology and realize the personalized recommendation service. In the final part of this paper, it examines the validity and predictability of the system through certain experiments and proves to be effective for library recommendation service.

关 键 词： 数据挖掘 关联规则 协同过滤 图书推荐

领　　域： [自动化与计算机技术] [自动化与计算机技术]