An Incremental Approach to Share-Frequent Itemsets Mining

Chayanan Nawapornanan, Sarun Intakosum, Veera Boonjing

Abstract

The share-frequent itemsets mining becomes an important topic in the mining of association rules because it can provide useful knowledge such as total quantity of items sold and total profit. In the past, the efficient MCShFI algorithm was successfully proposed to discover complete share-frequent itemsets on a database. When the database is updated, the algorithm can obtain current complete share-frequent itemsets by using the batch approach - mining the whole updated database. To improve mining execution time, we propose a new incremental approach to the problem with the Fast Update (FUP) concept. It obtains the current result by mining only new transactions and updating the previous existing result with this mined result.