基于FCA挖掘重述数据集的非冗余规则

Non-redundant Rules for Redescription Datasets Mining Based on FCA

以FCA为理论基础,应用Galois联络性质及其闭包运算,提出精确重述规则生成基和条件重述规则特征基的概念以及生成算法,论证由此构造的重述规则是最小非冗余的结论,设计最小非冗余重述规则挖掘的NRRM算法,并进行仿真计算和计算复杂性分析。

The notions and algorithms of generating basis for exact rules and the proper basis for conditional rules of redescription database are presented using operations of formal concept analysis and closure operator of Galois connection in the paper. It is true that constructed rules of redescription database are minimal non-redundant. And a new algorithm, called NRRM, for to discover minimal non-redundant and conditional rules is addressed, and simulation computation and computation complexity of NRRM algorithm are analyzed.