Incremental mining and re-mining of frequent patterns without storage of intermediate patterns

Data mining has been pervasively used for extracting business intelligence to support business decisionmaking processes. One of the most fundamental and important tasks of data mining is the mining of frequent patterns. When the transaction database is dynamic with data being updated constantly, incremental techniques must be used. Most techniques, though, adopt the "eager mining" approach that maintains a huge amount of intermediate patterns or data structures, which incurs expensive computational costs and consumes a lot of memory. Here an alternative "lazy mining" approach, called FP- impromptu, is proposed for incremental mining and re-mining of frequent patterns without storing intermediate patterns or massive data structures. Other possible applications and related issues of this approach are also discussed.