Abstract: During the clustering for big scale data, the traditional serial model has limitations, can not obtain a satisfied results within significant intervals. This paper proposes a concurrent clustering model based on MapReduce architechture under the Hadoop platform. The experiment result shows this model has a perfect linear speedup, and it outperforms the traditional clusting model, especially in dealing with massive data set.

Key words: data mining, algorithm, cloud computing, MapReduce, Hadoop