Advanced Pruning Strategies to Speed Up Mining Closed Molecular Fragments

In recent years several algorithms for mining frequent subgraphs in graph databases have been proposed, with a major application area being the discovery of frequent substructures of biomolecules. Unfortunately, most of these algorithms still struggle with fairly long execution times if larger substructures or molecular fragments are desired. In this paper we describe two advanced pruning strategies equivalent sibling pruning and perfect extension pruning that can be used to speed up the MoFa algorithm (introduced in [2]) in the search for closed molecular fragments, as we demonstrate with experiments on the NCI s HIV database