Incremental Association Mining using a Closed-Set Lattice
Aaron Ceglar, John F. Roddick, Aaron Ceglar, John F. Roddick · 2008
The use of closed-set algorithms to generate condensed accurate representations of a dataset’s frequent itemsets has been well documented. This paper presents a novel approach to incremental association mining in which the maintenance of the set of frequent itemsets is based upon the evolution of a closed-set lattice. This approach also creates a closed-set representation of the increment dataset, providing the user with insight to the increment’s effect upon the maintained lattice and provides an effective means of incorporating windowing functionality.