Parallel Analysis on Clustering Algorithm Based on Hadoop Cloud Computing Platform

Ordinary single mode clustering algorithm has been unable to adapt to the current actual information processing requirements both in efficiency or computing complexity theory, and the clustering algorithm based on Hadoop cloud computing platform parallel optimization analysis will indicate the trend clearer. Hadoop is actually a kind of open source project based on the type covered by the Apache, and it is also the basic operation mode of cloud computing distributed platform. Based on Hadoop cloud computing platform, the distributed file system (HDFS) can be used to realize data storage, and Map Reduce programming parallel optimization can be used for huge amounts of data information. According to the characteristics of the general clustering algorithms, combine it with Map Reduce programming mode, it can ensure programming developers to implement parallel need for insight into the structure optimized communication process, which is the parallel optimization of clustering algorithm stated in this paper. In view of this, the following will be based on general K-means algorithm in the initial clustering center selects random features and computing results limitations shall be optimized and improved, further approach based on Hadoop cloud computing platform to realize the related application of actual project.