abitofalchemy / GraMi

GraMi is a novel framework for frequent subgraph mining in a single large graph, GraMi outperforms existing techniques by 2 orders of magnitudes. GraMi supports finding frequent subgraphs as well as frequent patterns, Compared to subgraphs, patterns offer a more powerful version of matching that captures transitive interactions between graph nodes (like friend of a friend) which are very common in modern applications. Also, GraMi supports user-defined structural and semantic constraints over the results, as well as approximate results. For more details, check our paper: Mohammed Elseidy, Ehab Abdelhamid, Spiros Skiadopoulos, and Panos Kalnis. GRAMI: Frequent Subgraph and Pattern Mining in a Single Large Graph. PVLDB, 7(7):517-528, 2014.