Seminar on Data Ware Housing and Mining : mining association can often lead to discomfort many frequent item-sets and association rules. Recent studies have suggested an interesting alternative: mining frequent closed item-sets and their corresponding rules, which has the same strength as the association mining but substantially reduces the number of rules is presented.
CLOSET an efficient algorithm for closed item-set mining, which is developing three techniques:
(1) the application is often packaged tree model FP-tree mining closed item-sets without candidate generation,
(2) Development of a path compression technique to rapidly identify unique prefix patterns often closed, and
(3) Exploration of a partitioning mechanism of projection bases for mining databases and large scalable. Closet lager databases is efficient and scalable and is faster than previously proposed methods.
No comments:
Post a Comment