Automatic K selection method for the K — Means algorithm
Xiuli Shao, Huichao Lee, Yiwei Liu, Bo Shen · 2017
The key problem of the K-means clustering algorithm is to ascertain the appropriate number of clusters, that is K value, usually by ceaseless experiments to obtain it, the voting mechanism method was introduced in this paper to automatically determine the optimal number of clusters in a dataset. The voting mechanism method was introduced in this paper to automatically determine the optimal number of clusters in a dataset. In this method, the symbol K is used to represent the number of clusters, and a series candidate sets of K were obtained by using different validity indexes. Then, the frequency of each K was calculated, and we chose the highest frequency as the final number of clusters, that is the value of K. Using the method proposed in this paper to divide the real dataset of the ship tonnage, the experimental result is basically consistent with the rules of the actual operation, so the experimental verify the reliability and effectiveness of the approach presented in this paper.