Bayesian Rose Trees

Charles Blundell, Yee Whye Teh, Katherine Heller
Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, PMLR R8:65-72, 2010.

Abstract

Hierarchical structure is ubiquitous in data across many domains. There are many hier- archical clustering methods, frequently used by domain experts, which strive to discover this structure. However, most of these meth- ods limit discoverable hierarchies to those with binary branching structure. This lim- itation, while computationally convenient, is often undesirable. In this paper we ex- plore a Bayesian hierarchical clustering algo- rithm that can produce trees with arbitrary branching structure at each node, known as rose trees. We interpret these trees as mixtures over partitions of a data set, and use a computationally efficient, greedy ag- glomerative algorithm to find the rose trees which have high marginal likelihood given the data. Lastly, we perform experiments which demonstrate that rose trees are better models of data than the typical binary trees returned by other hierarchical clustering algorithms.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR8-blundell10a, title = {{B}ayesian Rose Trees}, author = {Blundell, Charles and Teh, Yee Whye and Heller, Katherine}, booktitle = {Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence}, pages = {65--72}, year = {2010}, editor = {Grünwald, Peter and Spirtes, Peter}, volume = {R8}, series = {Proceedings of Machine Learning Research}, month = {08--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r8/main/assets/blundell10a/blundell10a.pdf}, url = {https://proceedings.mlr.press/r8/blundell10a.html}, abstract = {Hierarchical structure is ubiquitous in data across many domains. There are many hier- archical clustering methods, frequently used by domain experts, which strive to discover this structure. However, most of these meth- ods limit discoverable hierarchies to those with binary branching structure. This lim- itation, while computationally convenient, is often undesirable. In this paper we ex- plore a Bayesian hierarchical clustering algo- rithm that can produce trees with arbitrary branching structure at each node, known as rose trees. We interpret these trees as mixtures over partitions of a data set, and use a computationally efficient, greedy ag- glomerative algorithm to find the rose trees which have high marginal likelihood given the data. Lastly, we perform experiments which demonstrate that rose trees are better models of data than the typical binary trees returned by other hierarchical clustering algorithms.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Bayesian Rose Trees %A Charles Blundell %A Yee Whye Teh %A Katherine Heller %B Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2010 %E Peter Grünwald %E Peter Spirtes %F pmlr-vR8-blundell10a %I PMLR %P 65--72 %U https://proceedings.mlr.press/r8/blundell10a.html %V R8 %X Hierarchical structure is ubiquitous in data across many domains. There are many hier- archical clustering methods, frequently used by domain experts, which strive to discover this structure. However, most of these meth- ods limit discoverable hierarchies to those with binary branching structure. This lim- itation, while computationally convenient, is often undesirable. In this paper we ex- plore a Bayesian hierarchical clustering algo- rithm that can produce trees with arbitrary branching structure at each node, known as rose trees. We interpret these trees as mixtures over partitions of a data set, and use a computationally efficient, greedy ag- glomerative algorithm to find the rose trees which have high marginal likelihood given the data. Lastly, we perform experiments which demonstrate that rose trees are better models of data than the typical binary trees returned by other hierarchical clustering algorithms. %Z Reissued by PMLR on 04 October 2026.
APA
Blundell, C., Teh, Y.W. & Heller, K.. (2010). Bayesian Rose Trees. Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R8:65-72 Available from https://proceedings.mlr.press/r8/blundell10a.html. Reissued by PMLR on 04 October 2026.

Related Material