K-means parallel multi-relational clustering algorithm for spatial data
Download as PDF
DOI: 10.25236/csam.2019.068
Author(s)
Huang Suyu, Wang Song
Corresponding Author
Huang Suyu
Abstract
This paper is a study of parallel clustering algorithm K-means. Firstly, the design idea of K-means clustering algorithm on a single computer is introduced. Secondly, the design idea of K-means clustering algorithm in cluster environment is elaborated in detail. When K-means clustering algorithm is faced with massive data, the complexity of time and space has become the bottleneck of K-means clustering algorithm. On the basis of fully studying the traditional K-Means clustering algorithm, this paper presents the design idea of parallel K-Means clustering algorithm, and gives the estimation formula of its acceleration ratio. The correctness and validity of the algorithm are proved by experiments.
Keywords
Clustering Algorithm, Parallel, Relational, Spatial Data, Cluster