English

Cutoff for exact recovery of Gaussian mixture models

Statistics Theory 2020-12-01 v3 Data Structures and Algorithms Information Theory math.IT Probability Machine Learning Statistics Theory

Abstract

We determine the information-theoretic cutoff value on separation of cluster centers for exact recovery of cluster labels in a KK-component Gaussian mixture model with equal cluster sizes. Moreover, we show that a semidefinite programming (SDP) relaxation of the KK-means clustering method achieves such sharp threshold for exact recovery without assuming the symmetry of cluster centers.

Keywords

Cite

@article{arxiv.2001.01194,
  title  = {Cutoff for exact recovery of Gaussian mixture models},
  author = {Xiaohui Chen and Yun Yang},
  journal= {arXiv preprint arXiv:2001.01194},
  year   = {2020}
}