ML Anthology

Very Fast EM-Based Mixture Model Clustering Using Multiresolution Kd-Trees

Andrew W. Moore
NeurIPS 1998 pp. 543-549
/neurips/1998/moore1998neurips-very/

Abstract

Clust ering is impor ta nt in m any fields including m anufac tlll'ing , biol og~', fin ance , a nd astronomy. l\Iixturp models arp a popula r ap(cid:173) proach due to their st.atist.ical found a t.ions, and EM is a very pop(cid:173) ular l1wthocl for fillding mixture models. EM, however, requires lllany accesses of the dat a , a nd thus h as been dismissed as imprac(cid:173) t ical (e.g. [9]) for d ata mining of enormous dataset.s. We present a nt' \· algorit.hm, baspd on thp l1lultiresolution ~.'Cl- trees of [5] , which dramatically reelucps the cost of EtlI-baspd clusteriug , wit.h savings rising linearl:; wit.h the number of datapoints. Although prespnt.pd lwre for maximum likplihoocl estimation of Gaussian mixt.ure mod(cid:173) f'ls , it. is also applicable to non-(~aussian models (provided class densit.ies are monotonic in Mahalanobis dist.ance), mixed categori(cid:173) cal/ nUllwric clusters. anel Bayesian nwthocls such as Antoclass [1].

PDF NeurIPS Semantic Scholar

Cite

Text

Moore. "Very Fast EM-Based Mixture Model Clustering Using Multiresolution Kd-Trees." Neural Information Processing Systems, 1998.

Markdown

[Moore. "Very Fast EM-Based Mixture Model Clustering Using Multiresolution Kd-Trees." Neural Information Processing Systems, 1998.](https://mlanthology.org/neurips/1998/moore1998neurips-very/)

BibTeX

@inproceedings{moore1998neurips-very,
  title     = {{Very Fast EM-Based Mixture Model Clustering Using Multiresolution Kd-Trees}},
  author    = {Moore, Andrew W.},
  booktitle = {Neural Information Processing Systems},
  year      = {1998},
  pages     = {543-549},
  url       = {https://mlanthology.org/neurips/1998/moore1998neurips-very/}
}