Constructing complete FP-tree for incremental mining of frequent patterns in dynamic databases

Abstract

This paper proposes a novel approach that extends the FP-tree in two ways. First, the tree is maintained to include every attribute that occurs at least once in the database. This facilitates mining with different support values without constructing several FP-trees to satisfy the purpose. Second, the tree is manipulated in a unique way that reflects updates to the corresponding database by scanning only the updated portion, thereby reducing execution time in general. Test results on two datasets demonstrate the applicability, efficiency and effectiveness of the proposed approach. © Springer-Verlag Berlin Heidelberg 2006.