Feasible itemset distributions in data mining: theory and application