The share measure for itemsets provides useful information about numerical values associated with transaction items, that the support measure cannot. Finding share frequent itemsets is difficult because share frequency is not downward closed when it is defined in terms of the itemset as a whole. The Item Add-back and Combine All Counted algorithms do not rely on downward closure and thus, are able to find share frequent itemsets that have infrequent subsets. These heuristic algorithms predict which itemsets should be counted in the current pass using information available at no additional processing cost.
展开▼