Top-down mining of top-k frequent closed patterns from microarray datasets

Abstract

Mining frequent closed patterns from microarray datasets has attracted more attention. However, most previous studies needed users to specify a minimum support threshold. In practice, it is not easy for users to set an appropriate minimum support threshold and discover the interesting patterns from huge frequent closed patterns. In this paper, we proposed an alternative mining task that mines top-k frequent closed patterns of length no less than min_l from microarray datasets, where k is the desired number of frequent closed patterns to be mined. An efficient algorithm TBtop is developed adopting top-down breadth-first search strategy. Our performance study showed that the strategy was effective in pruning search space. And in most cases, the algorithm TBtop outperformed the algorithm CARPENTER.