Just Relax and Come Clustering! : A Convexification of k-Means Clustering
作者： Fredrik Lindsten , Lennart Ljung , Henrik Ohlsson
DOI:
关键词:
摘要: k-means clustering is a popular approach to clustering. It easy implement and intuitive but has the disadvantage of being sensitive initialization due an underlying nonconvex optimization problem. In this paper, we derive equivalent formulation The takes form L0-regularized least squares We then propose novel convex, relaxed, sum-of-norms regularized inherits many desired properties advantage independent initialization.
参考文章(0)