基于粒计算的多层次关联规则挖掘技术研究

摘要

Granular computing theory is a view of world and methodology to look at the objective world. Muhi-Level association Rule Mining based on Granular Computing--MLAR-Grc, which imports granular computing thought and expressed using muhi-Level binary code, can get granulation of all leaves nodes also can get its' parent granulation and multi itemsets using OR and AND operation which Simplify the method of solving frequent k- ItemSets. AR-Grc can reduce time complexity and Space complexity and accordingly improve the efficiency of rule mining.%粒计算理论是一种看待客观世界的世界观和方法论．基于粒计算的多层次关联规则挖掘，引入了粒计算思想，采用多层次化二进制编码表示，只须一次数据集扫描便可获得所有叶节点粒，并能够由子粒的“或”、“与”运算获得父粒和多项集，简化求频繁一项集的求取方法，降低了算法的时间复杂度和空间复杂度提高了算法的效率．