Research on Algorithms of Mining K Frequent Patterns under Hadoop

Now, in the era of massive data, how to efficiently perform associate knowledge discovery has become a hot research topic in data mining. The traditional association rule algorithms often produce a number of models and rules, which not only reduces the efficiency and the effect of data mining, but also hinders the development and wide application of data mining. For this problem, MapReduce distributed computing model of Hadoop is adopted to study and analyze the association rule mining algorithm, and the idea of data partitioning is integrated, then an improved K parallel frequent pattern mining algorithm is proposed in this paper. K frequent patterns are directly generated based on transaction data and the produced results are filtered to obtain mining target by MapReduce framework. The experimental results show that the algorithm has good performance in efficiency of mining and the node expansion.