挖掘关注的语言值关联规则

为了解决利用RFCM算法划分数量型属性，并通过组合语言值进行语言关联规则挖掘中出现的规则数量太多，以及难于获得用户真正关注的规则等问题，提出了一种改进的语言值关联规则挖掘算法。通过最大隶属原则将记录在数量型属性上的取值转换为语言值，然后转换成布尔型属性关联规则挖掘问题。同时，给出一个能够度量语言值关联则简洁性和新奇性关注程度（兴趣度）的计算函数，用于减少选取关注语言值关联规则的工作量。采用本文提出的方法对一组实例数据进行实验，得到了关注程度较高的语言值关联规则。所采用的方法能适用于含有大量数量型属性的数据库，并能有效地获取用户关注的规则。

Mining Interesting Linguistic Value Association Rules

In order to resolve the problems that too much rules exist and great difficulties are involved in getting truly interesting rules, in mining linguistic value association rules and applying RFCM to partition quantitative attributes and linguistic values combination, the improved algorithm for mining linguistic value association is provided. In this algorithm, the values of quantitative attributes are transferred to linguistic values by maximum membership principle and then the problem of mining quantitative association rules is transferred to the problem of mining boolean association rules. At the same time, a new objective interestingness function, which can measure the novelty and simplicity of linguistic value association rules, is given to reduce workload of selecting the interesting linguistic value association rules. An experiment on a set of real world data is made and the linguistic value association rules with higher interestingness are obtained. The method adopted in this paper is applicable to database with many quantitative attributes and can effectively obtain the interesting linguistic value association rules.