A new share frequent itemsets mining using incremental BitTable knowledge

The share measure has been proposed as an important measure for mining association rules. The value of share itemsets provides useful information such as total profits and total customer purchased quantities associated with itemsets in database. The share-frequent itemsets mining problems become a very important research issue in data mining. Existing share-frequent itemsets mining algorithms are based on static database so knowledge must be rebuilded when the minimum share threshold is changed or database is modified either appended or updated. This paper proposes a novel BitTable knowledge for incremental and interactive share-frequent itemsets mining in multiple minimum share thresholds without rebuilding BitTable knowledge. It is effective for incremental and interactive mining to take advantage of the previous BitTable knowledge and the previous mining results. © 2011 AICIT.