基于粗糙集的关联规则挖掘方法的研究与应用

基于粗糙集理论知识，对关联规则挖掘算法作出一定的改进。该算法的主要思想是把集合的近似质量作为迭代准则，初始约简集是所有的条件属性集合，在保证近似质量不变的前提下通过逐步缩减的方式来求取约简集，保证了所求的约简不会减弱对问题的分类决策能力。约简后得到新的决策表，在此基础上应用基于贪心思想的Apriori算法挖掘关联规则。算法的主要优势是在不影响对问题分类决策能力的前提下，以较小的属性和候选项集数目以及有限的扫描次数生成决策规则。通过应用实例和实验分析验证了算法的有效性。

Research and application of association rules based on rough set

In this paper ,we propose anew association rule mining algorithm based on rough set .The main idea of the algorithm is to set the approximate quality of the collection as the iteration criterion ,and all the set of conditional attributes as the initial reduction .In the premise of ensuring the quality of the approximation ,the reduction set can be obtained by the method of gradual reduction ,which ensures the ability of classification and decision through the gradual reduction of the set .After the reduction ,the new decision table is obtained ,and then the improved Apriori algorithm based on greedy algorithm is applied to mining association rules .The main advantage of the method is that the decision rules are generated by the small number of attributes ,candidate item sets and the limited scanner of decision table , which don’ t affect the ability of classification . The validity of the method is verified by the examples and the experimental analysis .