K-means聚类算法及改进

K-means Clustering Algorithm and Its Improvement

聚类分析是研究如何在没有训练的条件下把样本划分为若干类,即把相似度较大的对象划分成同类,相异度大的则划分为不同的类型。K-Means算法属于其中之一,笔者首先概述K-means算法,然后针对K-means聚类算法的实际情况而提出具体的改进策略,希望能够在改进之后提升算法的效率。

Clustering analysis is to study how to divide the sample into several classes without training.That is,the objects with large similarity are classified into the same kind,and the degree of dissimilarity is divided into different types.K-Means algorithm is one of them.The author first summarizes the K-means algorithm,and then puts forward the concrete improvement strategy for the actual situation of the K-means clustering algorithm,hoping to improve the efficiency of the algorithm after the improvement.