GenMax with Index Technique for Pruning Bounded Frequent Itemsets

Mining frequent itemsets is one of the essential problems found in most of the data mining applications like extraction of association rules, correlations, multidimensional patterns, and also in some of the pattern matching tasks. Fast implementation and efficient utilization of memory for giving a solution to the problems involving frequent itemsets are highly required in transactional databases. GenMax, an algorithm which is mainly a search based is used for mining only the maximal frequent itemsets. It involves many optimization techniques to prune the original search space. A progressive focusing technique is applied here to perform maximal checking. Differential set propagation is used to perform fast frequency computation. But, the GenMax algorithm was not implemented with closed frequent itemset.
To handle this issue an innovative GenMax with index Technique is presented here for quick and effective pruning of Bounded Frequent Itemsets and thereby enumerate all maximal frequent itemsets and closed frequent itemsets. The Experimental results show better scalability of improved GenMax with incremental update strategy. To evaluate the performance a comparison is made between the proposed index oriented GenMax and existing GenMax for efficient pruning of the bounded frequent Itemsets in terms of item precision and also speed.