Abstract

Abstract: In many real-life applications, frequent pattern mining is widely employed. Since its debut, many academics have been attracted by the mining of common patterns from exact data. More emphasis has been paid in recent years to the mining of ambiguous data. Items in each transaction of this uncertain data are generally linked to existential probabilities that describe the probability that these articles will be present during the transaction. Compared to the accurate data extraction, the space for the search/solution of the uncertain data is significantly bigger owing to the existential probability. The models provided are based on the widely known Apriori and MapReduce algorithms. The algorithms suggested are split into three primary classes. Two Apriori MapReduce and AprioriMR methods are intended to properly extract patterns in big datasets. These algorithms extract any existing data items irrespective of their frequency. Tape the search space with the antimonotonic characteristic. Two more space trimming methods AprioriMR and top AprioriMR are introduced with the objective of identifying any common data patterns. Maximum common patterns. In addition, we live in the Big Data age. Furthermore, we offer certain improvements to increase its performance further. Experimental findings indicate the efficacy of the MapReduce for Big Data Analytics algorithm and its improvements in mining frequent patterns from unspecified data.Keywords: MapReduce (MR); Hadoop Archives (HAR); Sequential Pattern Mining (SPM); Parallel Frequent Pattern Growth(PFPG)