English

Quantizing Multiple Sources to a Common Cluster Center: An Asymptotic Analysis

Machine Learning 2020-10-26 v1 Information Theory math.IT Machine Learning

Abstract

We consider quantizing an LdLd-dimensional sample, which is obtained by concatenating LL vectors from datasets of dd-dimensional vectors, to a dd-dimensional cluster center. The distortion measure is the weighted sum of rrth powers of the distances between the cluster center and the samples. For L=1L=1, one recovers the ordinary center based clustering formulation. The general case L>1L>1 appears when one wishes to cluster a dataset through LL noisy observations of each of its members. We find a formula for the average distortion performance in the asymptotic regime where the number of cluster centers are large. We also provide an algorithm to numerically optimize the cluster centers and verify our analytical results on real and artificial datasets. In terms of faithfulness to the original (noiseless) dataset, our clustering approach outperforms the naive approach that relies on quantizing the LdLd-dimensional noisy observation vectors to LdLd-dimensional centers.

Keywords

Cite

@article{arxiv.2010.12546,
  title  = {Quantizing Multiple Sources to a Common Cluster Center: An Asymptotic Analysis},
  author = {Erdem Koyuncu},
  journal= {arXiv preprint arXiv:2010.12546},
  year   = {2020}
}
R2 v1 2026-06-23T19:35:55.998Z