The hardness of k-means clustering
暂无分享,去创建一个
We show that k-means clustering is an NP-hard optimization problem, even if k is fixed to 2.
[1] I. J. Schoenberg,et al. Metric spaces and positive definite functions , 1938 .
[2] Alan M. Frieze,et al. Clustering Large Graphs via the Singular Value Decomposition , 2004, Machine Learning.