Optimization and Parallelization of Fuzzy Clustering Algorithm Based on the Improved Kmeans++ Clustering
Yu Ma, Wenjuan Cheng · IOP Conference Series Materials Science and Engineering · 2020
Abstract In the field of big data, fuzzy clustering algorithm is one of the most widely used clustering algorithm. Although fuzzy c-means (FCM) algorithm performs well, it still is difficult to ascertain the number of clusters and is sensitive to initial clustering center. To solve these problems, an improved fuzzy clustering algorithm based on Spark is proposed in this article. The proposed algorithm integrates the L2 norm and uses the kmeans++ algorithm improved by the Canopy algorithm to initialize the cluster center. Experimental results show that the proposed algorithm performs well in clustering accuracy and computational performance.