K-means clustering algorithm: a brief review
<p>Bao Chong</p>
Abstract
K-means clustering is a very classical clustering algorithm, and it is also one of the representatives of unsupervised learning. It has the advantages of a simple idea, high efficiency, and easy implementation, so it is widely used in many fields. However, K-means clustering also has some limitations, such as the number of clusters, the value of K is challenging to select, the selection of initial class center, the detection of outliers, and so on. This paper introduces the traditional K-means clustering algorithm and its improved method in detail. The advantages and disadvantages of the improved algorithm are analyzed, and the existing problems are pointed out. The development direction and trend of the K-means algorithm have been prospected.