Ordering Effects in Clustering

Abstract

Incremental systems like Cobweb suffer from ordering effects: the clusters that they discover may differ with the presentation of objects. We introduce methods that retain Cobweb's probabilistic bias, which are efficient, but which are nonincremental and order independent, and thus more appropriate to certain data analysis tasks.

Cite

Text

Fisher et al. "Ordering Effects in Clustering." International Conference on Machine Learning, 1992. doi:10.1016/B978-1-55860-247-2.50026-7

Markdown

[Fisher et al. "Ordering Effects in Clustering." International Conference on Machine Learning, 1992.](https://mlanthology.org/icml/1992/fisher1992icml-ordering/) doi:10.1016/B978-1-55860-247-2.50026-7

BibTeX

@inproceedings{fisher1992icml-ordering,
  title     = {{Ordering Effects in Clustering}},
  author    = {Fisher, Douglas H. and Xu, Ling and Zard, Nazih},
  booktitle = {International Conference on Machine Learning},
  year      = {1992},
  pages     = {162-168},
  doi       = {10.1016/B978-1-55860-247-2.50026-7},
  url       = {https://mlanthology.org/icml/1992/fisher1992icml-ordering/}
}