An Efficient Approach for Parallel and Incremental Mining of Frequent Pattern in Transactional Database

In this paper, we provide an overview of parallel incremental association rule mining, which is one of the imminent ideas in the new and rapidly emerging research area of data mining. A useful tool for discovering frequently co-occurrent items is frequent itemset mining (FIM). Since its commencement, a number of significant FIM algorithms have been build up to increase mining performance. But when thedataset size is huge, both the computational cost and memory use can be toocostly. In this paper,we put frontward parallelizing the FP-Growth algorithm.We use MapReduce to execute the parallelization of FP-Growth algorithm. Henceforth, it splits the mining task into number of sub-tasks, implements these sub-tasks in parallel on nodes and then combines the results back for the final result.Experiments show that the result increases the computational speed as compared to apriori and fp-growth