ERA

Download the full-sized PDF of Bregman Divergence Clustering: A Convex ApproachDownload the full-sized PDF

Analytics

Share

Permanent link (DOI): https://doi.org/10.7939/R3R785W2T

Download

Export to: EndNote  |  Zotero  |  Mendeley

Communities

This file is in the following communities:

Graduate Studies and Research, Faculty of

Collections

This file is in the following collections:

Theses and Dissertations

Bregman Divergence Clustering: A Convex Approach Open Access

Descriptions

Other title
Subject/Keyword
convex
clustering
Bregman divergence
Type of item
Thesis
Degree grantor
University of Alberta
Author or creator
Cheng, Hao
Supervisor and department
Schuurmans, Dale (Computing Science)
Szepesvári, Csaba (Computing Science)
Examining committee member and department
Szepesvári, Csaba (Computing Science)
Schuurmans, Dale (Computing Science)
Boulanger, Pierre (Computing Science)
Department
Department of Computing Science
Specialization

Date accepted
2013-09-25T13:42:22Z
Graduation date
2013-11
Degree
Master of Science
Degree level
Master's
Abstract
Due to its wide application in various fields, clustering, as a fundamental unsupervised learning problem, has been intensively investigated over the past few decades. Unfortunately, standard clustering formulations are known to be computationally intractable. Although many convex relaxations of clustering have been proposed to overcome the challenge of computational intractability, current formulations of clustering remain largely restricted to spherical Gaussian or discriminative models and are susceptible to imbalanced clusters. To address these shortcomings, we propose a new class of convex relaxations that can be flexibly applied to more general forms of Bregman divergence clustering. By basing these new formulations on normalized equivalence relation matrix, we retain additional control on relaxation quality, which allows improvement in clustering quality. We fur- thermore develop optimization methods that improve scalability by exploiting recent implicit matrix norm methods. We find that the new formulations are able to efficiently produce tighter clusterings that improve the accuracy of state of the art methods.
Language
English
DOI
doi:10.7939/R3R785W2T
Rights
Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.
Citation for previous publication

File Details

Date Uploaded
Date Modified
2014-05-01T03:44:31.969+00:00
Audit Status
Audits have not yet been run on this file.
Characterization
File format: pdf (Portable Document Format)
Mime type: application/pdf
File size: 404843
Last modified: 2015:10:12 20:59:28-06:00
Filename: HaoCheng_Thesis_FinalToSubmit.pdf
Original checksum: bb20083c0afb83a176812821da71b2d9
Well formed: true
Valid: true
Page count: 67
Activity of users you follow
User Activity Date