基于SOM神经网络的高校图书馆个性化需求挖掘系统研究
|
摘要
借助SOM神经网络聚类算法无参数、精准度高的特点,本文对山西大学图书馆的Web访问行为进行聚类和优化分析。将聚类行为分为指数函数粗调整和线性函数微调整两个优化阶段,有效提升聚类速率和聚类效果。基于对用户分析结果的输出,将用户个人特征信息、用户行为数据以及文献数据库进行筛选整合,形成可靠性和可用性更高的关联数据集,并结合语义检索和属性值匹配等技术建构用户个性化服务推荐系统,进行有效性验证,实现图书馆内部主题推荐、图书推荐和专家推荐三个子系统的协同。
According to the characteristics of high precision and no parameter of the SOM neural network clustering algorithm, the paper, taking the web access behaviors of users in Shanxi University Library as an example, carried on optimized cluster analysis. The progress of clustering behavior could be divided into two stages, the rough adjustment training and the micro-adjustment training, which could improve the clustering rate and effect. Based on the output of analysis results, screening and integrating the user's personal characteristic information, users' behavior data and literature database, to linked data set reliable and available highly. And combining with the semantic retrieval and attributing matching technology, the user personalized service recommendation system was formed and verified effective. It realized the coordination among internal subjects recommending, books recommending and experts recommending.
According to the characteristics of high precision and no parameter of the SOM neural network clustering algorithm, the paper, taking the web access behaviors of users in Shanxi University Library as an example, carried on optimized cluster analysis. The progress of clustering behavior could be divided into two stages, the rough adjustment training and the micro-adjustment training, which could improve the clustering rate and effect. Based on the output of analysis results, screening and integrating the user's personal characteristic information, users' behavior data and literature database, to linked data set reliable and available highly. And combining with the semantic retrieval and attributing matching technology, the user personalized service recommendation system was formed and verified effective. It realized the coordination among internal subjects recommending, books recommending and experts recommending.
引文
[1]刘恩涛,李国俊,邱小花,等.MOOCs对高校图书馆的影响研究[J].图书馆杂志,2014,33(2):67-71.
[2]朱晓云.Web数据挖掘与个性化信息服务中用户研究[J].情报技术,2004(2):34-35.
[3]张莉萍.论图书馆个性化服务中用户信息的Web数据挖掘[J].情报资料工作,2007(2):101-104.
[4]欧阳烽.Web数据挖掘与高校数字图书馆的个性化服务[J].现代情报,2008,28(1):103-104.
[5]项冰冰,钱光超.聚类算法研究综述[J].电脑知识与技术(学术交流),2007,2(12):30-31.
[6]THAMARAISELVI G,KALIAMMAL A.Datamining:conceptsand techniques[M].[S.1.]:Morgan Kaufmann Publishers Inc.,2006.
[7]张德丰.MATLAB神经网络应用技术[M].北京:机械工业出版社,2012.
[8]徐涌,陈恩红,王熙法.基于神经网络的Web用户进行聚类分析[J].小型微型计算机系统,2001,22(6):700-702.
[9]段隆振,朱敏,王靓明.基于双Kohonen神经网络的Web用户访问模式挖掘算法[J].计算机工程与科学,2009,31(9):95-98.
[10]李赛,邹丽华.人工神经网络在聚类分析中的运用[J].经济视野,2016(15):238-239.
[11]柳胜国.数字图书馆个性化服务与Web日志挖掘数据预处理技术[J].现代情报,2007,27(7):65-67.
[12]黄敏,赖茂生.语义检索研究综述[J].图书情报工作,2008,52(6):63-66.
[13]HIENERT D,ZAPILKO B,SCHAER P,et al.Vizgr:linking data in visualizations[M].[S.1.]:Springer Berlin Heidelberg,2011.
[14]高劲松,周习曼,梁艳琪.面向关联数据的实体链接发现方法研究[J].中国图书馆学报,2016,42(6):85-101.
[2]朱晓云.Web数据挖掘与个性化信息服务中用户研究[J].情报技术,2004(2):34-35.
[3]张莉萍.论图书馆个性化服务中用户信息的Web数据挖掘[J].情报资料工作,2007(2):101-104.
[4]欧阳烽.Web数据挖掘与高校数字图书馆的个性化服务[J].现代情报,2008,28(1):103-104.
[5]项冰冰,钱光超.聚类算法研究综述[J].电脑知识与技术(学术交流),2007,2(12):30-31.
[6]THAMARAISELVI G,KALIAMMAL A.Datamining:conceptsand techniques[M].[S.1.]:Morgan Kaufmann Publishers Inc.,2006.
[7]张德丰.MATLAB神经网络应用技术[M].北京:机械工业出版社,2012.
[8]徐涌,陈恩红,王熙法.基于神经网络的Web用户进行聚类分析[J].小型微型计算机系统,2001,22(6):700-702.
[9]段隆振,朱敏,王靓明.基于双Kohonen神经网络的Web用户访问模式挖掘算法[J].计算机工程与科学,2009,31(9):95-98.
[10]李赛,邹丽华.人工神经网络在聚类分析中的运用[J].经济视野,2016(15):238-239.
[11]柳胜国.数字图书馆个性化服务与Web日志挖掘数据预处理技术[J].现代情报,2007,27(7):65-67.
[12]黄敏,赖茂生.语义检索研究综述[J].图书情报工作,2008,52(6):63-66.
[13]HIENERT D,ZAPILKO B,SCHAER P,et al.Vizgr:linking data in visualizations[M].[S.1.]:Springer Berlin Heidelberg,2011.
[14]高劲松,周习曼,梁艳琪.面向关联数据的实体链接发现方法研究[J].中国图书馆学报,2016,42(6):85-101.