Research and improvement of k-means parallel multi-association clustering algorithm

In this paper, k-means parallel clustering algorithm is studied. Firstly, this paper introduces the purpose and significance of k-means clustering algorithm. Secondly, we describe the process of clustering analysis, six classical clustering algorithms, the composition and operation of Hadoop Environment, and the K-MEANS algorithm in the cluster environment. In the environment of large data, the time and space complexity of k-means algorithm becomes an obstacle of k-means algorithm. Based on the research of a lot of traditional k-means algorithms, a parallel k-means algorithm is proposed, and the formula of its speed-increasing ratio is given. Experimental results show that the algorithm is correct and effective, and has a good clustering effect.