ExAnte: A processing method for frequent-pattern mining

The Knowledge Discovery and Delivery Laboratory in Pisa, has developed ExAnte, a simple yet effective preprocessing technique for frequent pattern mining. ExAnte exploits constraints to dramatically reduce the analyzed data to that containing patterns of interest. This data reduction, in turn, induces a strong reduction of the candidate pattern's search space, thus supporting substantial performance improvements in subsequent mining. The ExAnte preprocessing approach overcomes the suspected incompatibility between antimonotonicity and monotonicity. It uses the synergy of these two components to reduce data and search-space size. ExAnte's powerful underlying idea opens new application scenarios for constraint-based pattern discovery.