基于Hadoop的K-Means聚类算法优化与实现

ombined with hadoop cloud computing platform and MapReduce parallel programming framework, refered from K-Medoids clustering algorithm not sensitive to the outlier data, knowing the traditional K-Means clustering algorithm not suitable for large mass of data mining, this paper proposes an improved parallel K-Means clustering algorithm based on hadoop and name as HK-Means clustering algorithm. Design Map function to calculate the distance of each data record to each clustering center point and make them belong to one. Design Reduce function to update the clustering center points. Through the experiment, prove that compared with the traditional serial algorithm, HK-Means algorithm can indeed reduce the time complexity and also has good stability.