基于记录对比的关联规则挖掘算法设计

介绍了数据挖掘及关联规则的基本概念、数据挖掘算法设计的基本方法、现在流行的基于“支持-可信度”的经典关联规则挖掘算法，有针对性地指出了基于“支持度-可信度”的关联规则挖掘算法的诸多不足之处，在此基础上提出了基于记录对比的关联规则算法设计思想。该算法不再认为数据源中的备条记录是独立的，而认为可以通过记录对比，从记录与记录的差异中，找出某种潜在的关系。最后对基于记录对比的关联规则挖掘算法和基于“支持度-可信度”的关联规则挖掘算法进行了对比分析。

Designing Association-Rule Algorithms on Data Mining Based on the Relation of the Records

This paper introduces the perception of Data Mining and Association rule, the way of designing Data Mining Algorithms, and some shortcomings of the classical algorithms. Then, it comes up with the new algorithms of Data Mining based on the relation of the records of the data source. Working with this algorithm, the information and association rules will always be considered among the relations of the recorded data source. Eventually, this new algorithm is compared with classical algorithms.