The Finding of frequent itemset in big data is an important task in data mining and knowledgediscovery. The exponential daily growth of data, called “Big Data”, mining frequent patterns from the hugevolumes of data has many challenges due to memory requirement, multiple data dimensions, heterogeneityof data and so on. The complexities related to mining frequent item-sets from a Big Data can be minimizedby using Modified FP-growth algorithm and parallelizing the mining task with Map Reduce framework inHadoop. In this paper, a modified FP-growth based on directed graph with Hadoop framework will reducethe execution time for the massive database and works efficiently on number of nodes (computers). Thealgorithm was tested, our experimental results demonstrated that the proposed algorithm could scale welland efficiently process large datasets. In addition, it achieves improvement in memory consumption to storefrequent patterns and time complexity.
From big data mining to technical sport reports: the case of inertial measurement units