基于MapReduce的FCM聚类集成算法

针对传统的聚类集成算法难以高效地处理海量数据的聚类分析问题，提出一种基于MapReduce的并行FCM聚类集成算法。算法利用随机初始聚心来获取具有差异化的聚类成员，通过建立聚类成员簇间OVERLAP矩阵来寻找逻辑等价簇，最后利用投票法共享聚类成员中数据对象的分类情况得出最终的聚类结果。实验证明，该算法具有良好的精确度，加速比和扩展性，具有处理较大规模数据集的能力。

FCM Clustering Ensemble Algorithm Based on MapReduce

In order to improve the clustering ensemble algorithm`s efficiency of dealing with large-scale data set, a parallel FCM clustering ensemble algorithm based on MapReduce is put forward. This algorithm uses random initial cluster centers to get clustering members that have differentiation, and it can establish an OVERLAP matrix among the clusters to find the logical equivalence clusters. The final clustering results are obtained by voting that shares the information of the classification of the data object in clustering members. The experimental results show that the parallel FCM clustering ensemble algorithm has good precision and the advantages of both high speedup and good scalability. As a result, this algorithm has the ability of dealing with large-scale data set.