An Efficient Technique Using Parallel MFP for Mining Big Data

There is a sudden belief in the scientists and researchers that integrating a classification techniques and association rule mining can deliver more productive and precise classifiers which gives preferable outcomes like accuracy, low processing time over conventional data mining systems. The traditional classification algorithms wind up awkward with regards to manage the substantial amounts of data, since this data may require expansive processing time and high setup configuration tools to bargain. Because of the advancement of complex systems administration drifts and appropriated innovation techniques, makes distributed computing a reality in the execution of association rules combined with classification techniques. Generally in Apriori algorithm methodology, scanning of the external storage over and again require high I/O load stack which realizes the low execution performance in the framework. For FP-Growth technique, the adequacy is rely upon the inward size of the memory since mining process is performed on the base of huge tree-shape data structure and the cost is high to scan this structure. In this model, considered the association rule mining algorithm called Maximal Frequent Pattern (MFP) and proposes a parallel MFP model dependent on MapReduce Framework, called as PMFP algorithm which can process the large dataset on parallel circulated condition with high exactness and low mapping/reducer time.