Mining Frequent Itemset Using Parallel Computing Apriori Algorithm

Abstract

Frequent itemset mining from a large transactional database is a very time consuming process. A famous frequent pattern mining algorithm is Apriori. Apriori algorithm generates a frequent itemsets in loop manner, one frequent item adds in itemsets per loop. Apriori algorithm required multiple times dataset scans for itemset generation therefore it is time consuming process. Sometime Apriori become a holdup for large transactional dataset because of the long running time of the algorithm. This paper presents an efficient scalable Multi-core processor parallel computing Apriori that reduce the execution time and increase performance. Java concurrency libraries package used for the multi-core utilization that is easy and simple implementation technique. Furthermore, we compare the performance of Apriori sequential and parallel computing on the basis of time and varying support count for various transactional datasets.