Distributed Association Rules Mining of Varying Data Partition Size Using Nodesets

Manoj Sethi
2020 International Journal of Advanced Trends in Computer Science and Engineering  
Association rule mining in the distributed database has become an important area of research, where frequent pattern or itemsets are found in a large distributed data of varied sizes stored at multiple sites. In recent years, new efficient data structures have been proposed for the data mining. A new algorithm named as QDFIN(Quick distributed frequent itemset mining using nodeset) is proposed in this paper which uses the efficient nodeset data structureto store the candidate itemsets locally at
more » ... each site and zero-first technique to balance the load and pruning to reduce the candidate sets. The algorithm is implemented and the speed performance is compared with PFIN and FDM using FP-Growth algorithms. Results shows that the proposed algorithm not only outperform other algorithms on varying size data partition but also on uniform distributed data on 4, 5 and 6 node setups.
doi:10.30534/ijatcse/2020/181942020 fatcat:aeo3c7t46natngkrr2pnuy6lma