基于数据挖掘的图书馆书目推荐服务的研究
发布时间：2018-07-24 11:46
【摘要】：互联网的迅猛发展,给人们的生活方式带来了强大的冲击,丰富便捷的信息获取方式引发了全球范围内的信息革命。在这样的背景下,大多数商业网站建立起商品推荐系统,为人们提供更加直观有效的服务,但是至今为止,推荐服务却没有在图书馆应用方面得到足够的重视。 本文以提高图书管理中图书推荐服务为目的,将数据挖掘技术引入到图书馆管理系统中,文章首先对比了国内外图书推荐系统研究状况,指出图书馆信息推荐服务应该分为几个方面,需要哪些技术支持,并且介绍了常用的推荐技术,对比它们的优势和不足,选择适合进行书目推荐的推荐技术。然后详细的介绍进行书目推荐的数据挖掘方法：聚类分析方法、关联规则分析方法、决策树分析方法,选取每种数据挖掘方法中最适合的算法。在关联规则分析方法中,对其算法Apriori进行改进,引入矩阵的思想,将基于事务数据库的字符串运算转化为基于矩阵的布尔值运算,减少了算法运行过程中对数据库的访问,释放了内存空间,提高算法运行效率。最后以中北大学图书馆数据库中的借阅记录为基础,利用clementine软件对其进行数据挖掘为书目推荐服务提供实例参考。 进行数据挖掘时,共分数据预处理、数据挖掘实施、挖掘结果分析及结论建议四步进行,在四个步骤中,数据挖掘实施是重点阶段。本文利用聚类分析、关联规则分析和决策树分析三种方法对借阅记录实施数据挖掘,聚类分析和关联规则分析是从读者角度对数据进行处理,而决策树分析是从图书种类角度对数据进行处理,得到对该图书感兴趣的读者群,然后根据读者是否满足该读者群的特征,判断是否应该向读者推荐这种图书。其中,引入决策树分析方法是图书推荐服务的首次尝试。
[Abstract]:The rapid development of the Internet has brought a strong impact to people's way of life, and the rich and convenient way of obtaining information has triggered the information revolution in the world. In this context, most commercial websites set up a commodity recommendation system to provide people with more intuitive and effective services, but up to now, the recommended services have not been paid enough attention to in the application of libraries. In order to improve the book recommendation service in the book management, this paper introduces the data mining technology into the library management system. Firstly, the paper compares the research status of the book recommendation system at home and abroad. This paper points out that the library information recommendation service should be divided into several aspects and what technical support is needed, and introduces the commonly used recommended technologies, compares their advantages and disadvantages, and selects the recommended technology suitable for bibliographic recommendation. Then it introduces the data mining methods of bibliographic recommendation in detail: cluster analysis method, association rule analysis method, decision tree analysis method, and select the most suitable algorithm in each data mining method. In the analysis method of association rules, the algorithm Apriori is improved, the idea of matrix is introduced, the string operation based on transaction database is transformed into Boolean operation based on matrix, and the access to database is reduced. The memory space is freed and the efficiency of the algorithm is improved. Finally, based on the borrowing records in the database of the Central North University Library, the author uses the clementine software to mine the data for the bibliographic recommendation service. Data mining is divided into four steps: data preprocessing, data mining implementation, mining result analysis and conclusion and suggestion. Among the four steps, data mining implementation is the key stage. In this paper, we use clustering analysis, association rule analysis and decision tree analysis to implement data mining for loan records. Cluster analysis and association rule analysis deal with data from the perspective of readers. The decision tree analysis is to process the data from the perspective of book type to get the readers who are interested in the book, and then judge whether the reader should recommend the book to the reader according to whether the reader satisfies the characteristics of the reader. Among them, the introduction of decision tree analysis method is the first attempt of book recommendation service.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP311.13
本文编号：2141276
[Abstract]:The rapid development of the Internet has brought a strong impact to people's way of life, and the rich and convenient way of obtaining information has triggered the information revolution in the world. In this context, most commercial websites set up a commodity recommendation system to provide people with more intuitive and effective services, but up to now, the recommended services have not been paid enough attention to in the application of libraries. In order to improve the book recommendation service in the book management, this paper introduces the data mining technology into the library management system. Firstly, the paper compares the research status of the book recommendation system at home and abroad. This paper points out that the library information recommendation service should be divided into several aspects and what technical support is needed, and introduces the commonly used recommended technologies, compares their advantages and disadvantages, and selects the recommended technology suitable for bibliographic recommendation. Then it introduces the data mining methods of bibliographic recommendation in detail: cluster analysis method, association rule analysis method, decision tree analysis method, and select the most suitable algorithm in each data mining method. In the analysis method of association rules, the algorithm Apriori is improved, the idea of matrix is introduced, the string operation based on transaction database is transformed into Boolean operation based on matrix, and the access to database is reduced. The memory space is freed and the efficiency of the algorithm is improved. Finally, based on the borrowing records in the database of the Central North University Library, the author uses the clementine software to mine the data for the bibliographic recommendation service. Data mining is divided into four steps: data preprocessing, data mining implementation, mining result analysis and conclusion and suggestion. Among the four steps, data mining implementation is the key stage. In this paper, we use clustering analysis, association rule analysis and decision tree analysis to implement data mining for loan records. Cluster analysis and association rule analysis deal with data from the perspective of readers. The decision tree analysis is to process the data from the perspective of book type to get the readers who are interested in the book, and then judge whether the reader should recommend the book to the reader according to whether the reader satisfies the characteristics of the reader. Among them, the introduction of decision tree analysis method is the first attempt of book recommendation service.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP311.13
【参考文献】
相关期刊论文 前10条
1 钱宏;;数据挖掘预处理技术的研究[J];电脑知识与技术;2010年17期
2 耿鑫;刘晋佩;;数据挖掘中的推荐算法综述[J];电脑知识与技术;2012年19期
3 吴兵;叶春明;;基于效用的个性化推荐方法[J];计算机工程;2012年04期
4 余力,刘鲁;电子商务个性化推荐研究[J];计算机集成制造系统;2004年10期
5 常勇生;;国内外数字图书馆个性化信息服务现状与建设趋势[J];科技情报开发与经济;2007年28期
6 陈锦;吴扬扬;;Apriori算法在高校图书馆图书推荐中的应用[J];河南科技学院学报(自然科学版);2012年04期
7 丁雪;;基于数据挖掘的图书智能推荐系统研究[J];情报理论与实践;2010年05期
8 黄月红;周秀梅;覃泽;;基于关联规则的图书借阅服务推荐方法[J];图书馆界;2010年04期
9 阴江烽;;面向科研学术对象服务的个性化图书推荐系统研究[J];探求;2012年04期
10 林郎碟;王灿辉;;Apriori算法在图书推荐服务中的应用与研究[J];计算机技术与发展;2011年05期
，本文编号：2141276
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2141276.html
