Exceptional Model Mining

In most databases, it is possible to identify small partitions of the data where the observeddistribution is notably different from that of the database as a whole. In classical subgroup discovery