摘要
已有的关联规则算法研究的对象通常是基于单一类型属性的关系表,算法的实际应用范围受到挖掘对象属性类型的限制。在对基于熵值离散法的量化关联规则进行改进的基础上,提出了面向混合属性关系表的关联规则挖掘思想,并对其应用于股票信息的关联规则分析做出了阐述。试验证明改进后的算法是可行的、有效的。今后还可对不同时间段、不同置信度、不同支持条件下的股票信息中蕴涵的关联规则进行研究,并进行相应的验证与修改。
Previous association rules proposed by other researchers mainly focus on single attributes database, which applications are restricted by the attributes type of mining objects. This paper will provide an introduction to the improved association rules for hybrid attributes based on entropy, and then expatiate on its applications. The experiments show that this approach is feasible and effective. In the future for the different time periods and different confidence level, supporting the different conditions of the stock of information contained in a study of the association rules, and the corresponding verification and revision.
出处
《金陵科技学院学报》
2008年第3期37-41,共5页
Journal of Jinling Institute of Technology
关键词
关联规则
熵
混合属性
association rules
entropy
hybrid attributes