Experimental Comparative Research on Clustering Algorithm Based on Spark Platform
Bo Zhang, Xinhang Li, Yukun Yang, Yanan Ma, Junjie Geng, Lei Chen · 2019
In the era of big data, data mining has become the focus of research in various fields. Cluster analysis is one of its core processing methods. A large number of iterative calculations make cluster analysis computation extremely time-consuming. At the same time, data mining based on Spark platform is not yet mature at home and abroad, and the research on clustering algorithm based on Spark is still very scarce. Based on the above background, this paper focuses on the Spark framework, and conducts research in-depth on it. Using the advantages of its elastic distribution data set and the characteristics of parallelization, this paper improves the processing speed of clustering analysis, and compares the clustering results of clustering algorithms - K-means, K-means++, PIC and GMM through experimental comparison.