Improving Performance Of Sequential Rule Mining With Parallel Computing

Abstract

Aiming to improve the performance of sequential rules mining algorithm for the large-scale data sets, this paper presents parallel algorithms for mining sequential rules which directly using MPJ Express for passing message base on multicore configuration and cluster configuration (master-slave structural model). Results analysis showed that the mining time of the parallel algorithms (both multicore and cluster model) which proposed in this paper have better performances compared with the sequential state-of-art algorithm.