基于项目序列集操作理论的关联规则的挖掘算法

Association Rules'Mining Architecture Based on Operating Theory of Itemsequences Set

在关联规划挖掘理论研究上,首次给出了项目序列集格空间,并且探讨了在这个空间上的基本操作算子、基于项目序列集格空间及其操作,建立了关联规则挖掘模型.在关联规则挖掘算法方面,设计了基于项目序列集操作理论的关联规则挖掘算法SIS,该算法执行时间整体上优于Apriori算法,而且随着数据量的增大,该算法执行时间的增长幅度也小于Apriori算法.

This paper puts forward set of itemsequences space under the study of assosiation rules, discusses the basis operator, and its operating in the space based on the set of itemsequences, estabilishes mining mode of assosiation rules. In association rule mining, we first define Set of Itemsequences and give some operators on this algebra lattice. Applying such theoretic results, we design an algorithm - SIS for mining association rules, which is more efficient with one pass to the database and without large candidates generated and stored than Apriori. With mining large - scale databases, it is a more smart strategy to reduce data capability than current one like Apriori.