Mining constrained cube gradient using condensed cube

Constrained cube gradient mining is an important mining task and has broad applications. The goal of constrained cube gradient mining is to extract the interesting pairs of gradient-probe cells from a data cube. The constrained cube gradient mining faces the obstacle of large requirements in time and space for generating combined gradient cells and probe cells. In this paper, we explore the condensed cube approach that is a novel and efficient data organization technique to the mining of constrained cube gradients. A new algorithm based on the condensed cube approach is developed through an extension of the existing efficient mining algorithm LiveSet. Results of the experiments show our algorithm is more effective than the existing algorithm on the performance of mining constrained cube gradient.