DHPTID-hybrid algorithm: A hybrid algorithm for association rule mining

Abstract

Direct Hashing and Pruning algorithm of ARM performs well at initial passes by smaller candidate 2-itemset generation and turns out to be very powerful to facilitate initial itemset generation. Efficient pruning technique of AprioriTid algorithm is highly effective for frequent itemset generation in the later passes. Theoretical and practical studies reveals the underneath facts for the scope of improvement for both the algorithms. This paper describes a hybrid approach for association rule mining, based on coupling of best part of two well known ARM algorithms, AprioriTid and DHP with a binary approach which significantly improves the performance. © 2010 Springer-Verlag.