Dirichlet Process Mixtures of Generalized Mallows Models

Marina Meila, Harr Chen
Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, PMLR R8:365-374, 2010.

Abstract

We present a Dirichlet process mixture model over discrete incomplete rankings and study two Gibbs sampling inference techniques for estimating posterior clusterings. The first ap- proach uses a slice sampling subcomponent for estimating cluster parameters. The sec- ond approach marginalizes out several cluster parameters by taking advantage of approx- imations to the conditional posteriors. We empirically demonstrate (1) the effectiveness of this approximation for improving conver- gence, (2) the benefits of the Dirichlet pro- cess model over alternative clustering tech- niques for ranked data, and (3) the applica- bility of the approach to exploring large real- world ranking datasets.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR8-meila10a, title = {{D}irichlet Process Mixtures of Generalized Mallows Models}, author = {Meila, Marina and Chen, Harr}, booktitle = {Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence}, pages = {365--374}, 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/meila10a/meila10a.pdf}, url = {https://proceedings.mlr.press/r8/meila10a.html}, abstract = {We present a Dirichlet process mixture model over discrete incomplete rankings and study two Gibbs sampling inference techniques for estimating posterior clusterings. The first ap- proach uses a slice sampling subcomponent for estimating cluster parameters. The sec- ond approach marginalizes out several cluster parameters by taking advantage of approx- imations to the conditional posteriors. We empirically demonstrate (1) the effectiveness of this approximation for improving conver- gence, (2) the benefits of the Dirichlet pro- cess model over alternative clustering tech- niques for ranked data, and (3) the applica- bility of the approach to exploring large real- world ranking datasets.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Dirichlet Process Mixtures of Generalized Mallows Models %A Marina Meila %A Harr Chen %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-meila10a %I PMLR %P 365--374 %U https://proceedings.mlr.press/r8/meila10a.html %V R8 %X We present a Dirichlet process mixture model over discrete incomplete rankings and study two Gibbs sampling inference techniques for estimating posterior clusterings. The first ap- proach uses a slice sampling subcomponent for estimating cluster parameters. The sec- ond approach marginalizes out several cluster parameters by taking advantage of approx- imations to the conditional posteriors. We empirically demonstrate (1) the effectiveness of this approximation for improving conver- gence, (2) the benefits of the Dirichlet pro- cess model over alternative clustering tech- niques for ranked data, and (3) the applica- bility of the approach to exploring large real- world ranking datasets. %Z Reissued by PMLR on 04 October 2026.
APA
Meila, M. & Chen, H.. (2010). Dirichlet Process Mixtures of Generalized Mallows Models. Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R8:365-374 Available from https://proceedings.mlr.press/r8/meila10a.html. Reissued by PMLR on 04 October 2026.

Related Material