Data mining in a large database environment

Abstract

Data mining, the process of discovering hidden and potentially useful information from very large databases, has been recognized as one of the most promising research topics in the 1990s. The essential problem faced in the mining of association rules is the generation of large items, which are items that are present in at least s% (minimal support) of the total database tuples. As the large items and their counts information usually require much storage space, the minimal cover concept is introduced to achieve reductions in the storage size. Percentage contour, an extension of minimal cover, is further introduced to aid in the handling of large databases.