Tree Projection-Based Frequent Itemset Mining on Multicore CPUs and GPUs

摘要: Frequent itemset mining (FIM) is a core operation for several data applications as association rules computation, correlations, document classification, and many others, which has been extensively studied over the last decades. Moreover, databases are becoming increasingly larger, thus requiring higher computing power to mine them in reasonable time. At same time, advances high performance platforms transforming into hierarchical parallel environments equipped with multi-core processors many-core accelerators, such GPUs. Thus, fully exploiting these systems perform FIM tasks poses challenging critical problem that we address this paper. We present efficient GPU accelerated parallelizations of Tree Projection, one most competitive algorithms. The experimental results show our Projection implementation scales almost linearly CPU shared-memory environment after careful optimizations, while versions up 173 times faster than standard version.