增量式关联规则维护的时间窗口技术

A Time Windowing Technique for the Incremental Maintenance of Association Rules

数据库中的知识发现是指在大型数据集中识别有效、新奇、潜在有用、且最终可理解模式的非平凡的过程。人们已经提出了许多种知识发现算法 ,然而 ,由于数据随时间变化而导致的所发现知识的更新维护问题却较少研究。笔者提出一种用于增量式关联规则维护的时间窗口技术。该技术可以集中在当前数据中发现强关联规则 ,避免利用过时数据。为了避免在已有数据上重新发现 ,降低数据存储开销 。

Knowledge discovery in databases is the nontrivial process of identifying valid, novel, potentially useful, and ultimately understandable patterns in large data set. Many techniques have been developed for knowledge discovery, however, comparatively little investigation has been made on maintenance and update of the discovered knowledge as the underlying data changes over time. We propose a time windowing technique for the incremental maintenance of association rules, which can focus on finding strong association rules within current data and avoid the use of outdated data. To avoid re-finding upon existing data and reduce data storage overhead, we store near strong association rules.