基于分类频繁模式挖掘的书目推荐策略与算法

Bibliographic Recommendation Strategy and Algorithm Based on Classified Frequency Pattern Mining

针对高校图书馆读者特点以及图书借阅的时间相关性等特征,提出一种基于分类频繁模式挖掘的图书推荐策略与算法。该策略根据高校图书馆图书借阅的时间周期性特征,将图书借阅数据划分为不同时间段的数据子集,进一步依据读者不同属性值,对不同类别读者在不同时段的数据子集上进行频繁模式挖掘。当某一个特定读者在特定时间进入系统时,便依据该作者所属类别在特定时段的频繁模式向该作者提供书目推荐服务。

For the characteristics of university libraries’readers and the relevant features of the time at which library’s books are borrowed,the bibliographic recommendation strategy and algorithm are pro posed based on classified frequency pattern mining.In terms of the characteristics of time periodicity for university’s borrowed books,the strategy divides book borrowing dates into different data subsets in differ ent time quantum.Further based on different readers’attribute value,frequency pattern mining is carried out for different categories of readers in a different subset of the data.When a specific reader enters the system at a particular time,based on the category of the author to the frequency patterns at a particular time,the system provides bibliographical recommendation service.