Beyond Hartigan Consistency: Merge Distortion Metric for Hierarchical Clustering

@inproceedings{Eldridge2015BeyondHC,
  title={Beyond Hartigan Consistency: Merge Distortion Metric for Hierarchical Clustering},
  author={Justin Eldridge and Mikhail Belkin and Yusu Wang},
  booktitle={COLT},
  year={2015}
}
Hierarchical clustering is a popular method for analyzing data which associates a tree to a dataset. Hartigan consistency has been used extensively as a framework to analyze such clustering algorithms from a statistical point of view. Still, as we show in the paper, a tree which is Hartigan consistent with a given density can look very different than the correct limit tree. Specifically, Hartigan consistency permits two types of undesirable configurations which we term over-segmentation and… CONTINUE READING
Highly Cited
This paper has 19 citations. REVIEW CITATIONS

From This Paper

Topics from this paper.

References

Publications referenced by this paper.
Showing 1-10 of 19 references

Consistency of Single Linkage for High-Density Clusters

  • J. A. Hartigan
  • Journal of the American Statistical Association,
  • 1981
Highly Influential
4 Excerpts

2013)) LetM be an m-dimensional Riemannian manifold with intrinsic metric dM. Given a subset A ⊆M and a parameter r >

  • Chazal
  • 2013

byNr(A) to be the minimum number of closed geodesic balls of radius r needed to cover A (the balls do not have to be centered in A)

  • Chazal
  • 2013

Similar Papers

Loading similar papers…