一种基于关联挖掘的缺省规则更新算法

为了解决缺省关联规则的增量挖掘问题,在算法DRMBAR的基础上,结合粗糙集理论及频繁模式树结构,提出了一种基于关联规则的缺省规则更新算法IADRBAR,该算法主要考虑最小支持度发生变化时缺省规则的更新问题,即在新的最小支持度下,如何高效地生成新的关联规则. IADRBAR在最坏的情况下仅须扫描决策表一遍,并利用上一次已经挖掘出的频繁项目集及关联规则,有效地提高缺省规则的更新效率.理论分析和实验结果表明算法是有效可行的.

Incremental algorithm based on association mining for default rules

For efficiently mining default rules, a novel algorithm based on association rules, DRMBAR (default rules mining based on association rules), is proposed using FP-tree (frequent pattern tree) structure. In this paper, incremental mining of default rules is addressing. Algorithm IADRBAR (incremental algorithm of default rules based on association rules) is proposed, which is based on association rules and the rough set theory. It mainly solves the problem of incremental updating of default rules when minimum support measure threshold is dynamically adjusted. In the worst case, IADRBAR only needs to scan the decision table once by utilizing the frequent item sets and association rules which have been obtained during the last round. Thus, IADRBAR can greatly improve incremental updating efficiency of default rules. Theoretical analysis and experimental results show that IADRBAR is efficient and effective.