English

Convergence of online $k$-means

Machine Learning 2022-02-23 v1

Abstract

We prove asymptotic convergence for a general class of kk-means algorithms performed over streaming data from a distribution: the centers asymptotically converge to the set of stationary points of the kk-means cost function. To do so, we show that online kk-means over a distribution can be interpreted as stochastic gradient descent with a stochastic learning rate schedule. Then, we prove convergence by extending techniques used in optimization literature to handle settings where center-specific learning rates may depend on the past trajectory of the centers.

Keywords

Cite

@article{arxiv.2202.10640,
  title  = {Convergence of online $k$-means},
  author = {Sanjoy Dasgupta and Gaurav Mahajan and Geelon So},
  journal= {arXiv preprint arXiv:2202.10640},
  year   = {2022}
}