Enhanced Index Based GenMax for Frequent Item Set Mining

In many data mining applications such as the discovery of association rules, strong rules, and many other important discovery tasks, mining frequent item sets is a fundamental and essential problem. Methods have been implemented for mining frequent item sets using a prefix-tree structure, for storing compressed information GenMax is used for mining maximal frequent item sets. It uses a technique called progressive focusing to perform maximal checking, and differential set propagation to perform fast frequency computation. Genmax algorithm was not implemented for closed frequent item set.
The proposal in this paper present an improved index based enhancement on Genmax algorithm for effective fast and less memory utilized pruning of maximal frequent item and closed frequent item sets. The extension induces a search tree on the set of frequent closed item sets thereby we can completely enumerate closed item sets without duplications. The memory use of mining the maximal frequent item set does not depend on the number of frequent closed item sets. The proposed model reduce the number of disk I/Os and make frequent item set mining scale to large transactional databases. Experimental results shows a comparison of improved index based GenMax and existing GenMax for efficient pruning of maximal frequent and closed frequent item sets in terms of item precision and fastness.