A Bitmap Approach for Closed and Maximal Frequent Itemset Mining

Abstract

Association Rule Mining (ARM) plays a fundamental role in many data mining tasks that attempt to find interesting patterns from databases, such as correlations, sequences, episodes, classifiers, clusters, etc. Frequent Itemset Mining (FIM) is one of the essential parts of ARM which is an active research area and a large number of algorithms have been developed. FIM algorithms may be depth-first or breadth-first approach. Most depth-first based approaches do not effectively address the cost of database projections. Consequently, their performance gets degraded severely as the total number of frequent itemsets in a database increases significantly. To solve this problem, a three - strategy adaptive algorithm, Bitmap Itemset Support Counting (BISC) with the closed and maximal frequent itemset is presented in this paper. The core strategy of the proposed algorithm is the usage of the closed and maximal frequent itemset detection which could be able to find least number of association rules. The proposed approach is compared with the conventional approach to evaluate the performance and accuracy. The experimental results show that the proposed approach outperforms the existing approach. Keywords: Data Mining Algorithms; Bitmap; Frequent Itemset Mining; Association Rule Mining.