【作者】 高雅；

【导师】 戴齐；

【作者基本信息】 西南交通大学 ， 计算机应用技术， 2004， 硕士

【摘要】 关联规则挖掘是数据挖掘领域中一个重要的研究方向。为了解决数量型关联规则挖掘过程中“边界划分过硬”的问题,人们将模糊集的有关概念引入到关联规则挖掘中,提出了“模糊关联规则”。本文利用模糊蕴涵算子定义蕴涵度提出了解决模糊关联规则挖掘问题的一种新方法,并就不同模糊蕴涵算子对关联规则挖掘结果的影响进行了讨论。主要工作如下: 1.首先利用模糊蕴涵算子定义了冗余模糊关联规则,分析了冗余模糊关联规则的性质,提出删除冗余模糊关联规则提高执行效率的新算法。并通过实例加以描述。 2.分析、比较了利用支持度和蕴涵度解决模糊关联规则挖掘问题的一些方法,讨论了选择不同算法的条件。 3.定义了模糊关联规则集间的距离公式,并分析了该距离公式的合理性. 4.用18种常用的模糊蕴涵算子分别定义蕴涵度,并利用上述距离公式对其挖掘出的18个强模糊关联规则集合进行了聚类分析,得出聚类结果,并讨论了采用不同蕴涵算子对规则挖掘结果的影响。更多还原

【Abstract】 Association rule mining is one of the key points in the research field of data mining. To deal with the problem of sharp boundary in mining quantative association rules, fuzzy sets has been introduced to data mining. Such kind of association rule mining is called mining fuzzy association rule. Using degree of implication defined on fuzzy implication operator, the present work propose a new method to mine fuzzy association rule, and discuss the infection of different fuzzy implication operators on it. Conceretly, the main work includes:1. Firstly, we analyzed some properties of fuzzy association rules and gave the definition of redundant fuzzy association rules (RFA) . Then, using degree of implication on fuzzy implication operator, we introduced a new algorithm to mine fuzzy association rules from frequent itemsets. At last, an example was given to illustrate our method.2. Secondly, we compared some general methods of solving the problem of fuzzy association rule mining by degree of support and degree of implication. And we discussed the different situations when certain method should be used.3. Thirdly, we gave the definition of distance between two fuzzy association rule sets, and anylized the rationality of this definition.4. Fourthly, using the distance formula mentioned above, we did some cluster analysis of the 18 fuzzy association rules obtained from the measure of degree of implication defined by the 18 frequently used fuzzy association operators.更多还原