This paper applied equivalence relation in rough set theory, combined refinement and generalization inrnknowledge system with searching for frequent itemset interatively layer by layer in Apriori, explored quantitative association rules in data mining, and put forward a new mining algorithm for quantitative association rules ( Mqars) . In the algorithm of Mqars , it didn't need to change quantitative attributes to boolean, reduced the non-candidate frequent itemsets, computed the support degree of itemset conveniently and proved the efficiency of the algorithm. In the end, this paper described the algorithm, its implementation, example and anlysis. In the end, it gave the example and compared algorithm Maqars with Maqa from aspect of time complexity and the efficiency . The result of analysis and comparison showed that the validity of the algorithm.%本文应用粗糙集理论中等价关系的概念,结合知识系统细化和泛化的思想以及Apriori算法中逐层搜索迭代求取频繁项集的思想,对数据挖掘中的多值属性关联规则问题进行研究,提出一种新的多值属性关联规则挖掘算法Mqars.Mqars的主要特点是无需将多值属性转化为布尔型属性,可以尽早地约简非候选的频繁项集,方便快捷地计算出项集支持度,提高多值属性关联规则挖掘效率.论文给出了Mqars算法详细描述、具体实现过程和算法实例及分析.最后设计实验环节对Mqars算法与传统的Maqa算法在时间复杂度和算法效率方面进行比对和分析,分析与比对的实验结果表明了该算法的有效性.
展开▼
