Enhancing FP-Growth Performance Using Multi-threading based on Comparative Study

Abstract

The time required for generating frequent patterns plays an important role in mining association rules, especially when there exist a large number of patterns and/or long patterns. Association rule mining has been focused as a major challenge within the field of data mining in research for over a decade. Although tremendous progress has been made, algorithms still need improvements since databases are growing larger and larger. In this research we present a performance comparison between two frequent pattern extraction algorithms implemented in Java, they are the Recursive Elimination (RElim) and FP-Growth, these algorithms are used in finding frequent itemsets in the transaction database. We found that FP-growth outperformed RElim in term of execution time. In this context, multithreading is used to enhance the time efficiency of FP-growth algorithm. The results showed that multithreaded FP-growth is more efficient compared to single threaded FP-growth.

Analysis of frequent itemset generation based on trie data structure in Apriori algorithm

“…From another perspective, it is common to use multi-thread and computing power with multicore architecture [12] in supporting data processing. This raises the suspicion that multi-thread in [13] and [14] can produce better performance also in getting this frequent itemset, as well as a challenge on how to determine the best performance process architecture that is applied to which subprocesses as threads and how big is the increase for the best multi-thread architecture in a single server environment, where an experiment in multi-node server environment was proposed by [15], [16].…”