<?xml version="1.0" encoding="UTF-8"?><?xml-stylesheet type="text/xsl" href="static/style.xsl"?><OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"><responseDate>2026-11-22T05:09:23Z</responseDate><request verb="GetRecord" identifier="oai:uwspace.uwaterloo.ca:10012/3934" metadataPrefix="dim">https://dspacemainprd01.lib.uwaterloo.ca/server/oai/request</request><GetRecord><record><header><identifier>oai:uwspace.uwaterloo.ca:10012/3934</identifier><datestamp>2017-03-21T13:56:34Z</datestamp><setSpec>com_10012_9900</setSpec><setSpec>com_10012_1</setSpec><setSpec>col_10012_9908</setSpec><setSpec>col_10012_6</setSpec></header><metadata><dim:dim xmlns:dim="http://www.dspace.org/xmlns/dspace/dim" xmlns:doc="http://www.lyncode.com/xoai" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.dspace.org/xmlns/dspace/dim http://www.dspace.org/schema/dim.xsd">
   <dim:field mdschema="dc" element="contributor" qualifier="author" authority="2766dcaa-416d-41cc-8f87-8f0459fd2c1d" confidence="600">Ayad, Hanan</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="accessioned">2008-08-29T18:47:14Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="available">2008-08-29T18:47:14Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="issued">2008-08-29T18:47:14Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="submitted">2008-08</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="uri">http://hdl.handle.net/10012/3934</dim:field>
   <dim:field mdschema="dc" element="description" qualifier="abstract" lang="en">Over the past few years, there has been a renewed interest in the consensus&#xd;
problem for ensembles of partitions. Recent work is primarily motivated by the&#xd;
developments in the area of combining multiple supervised learners. Unlike the&#xd;
consensus of supervised classifications, the consensus of data partitions is a&#xd;
challenging problem due to the lack of globally defined cluster labels and to&#xd;
the inherent difficulty of data clustering as an unsupervised learning problem.&#xd;
Moreover, the true number of clusters may be unknown.  A fundamental goal of&#xd;
consensus methods for partitions is to obtain an optimal summary of an ensemble&#xd;
and to discover a cluster structure with accuracy and robustness exceeding those&#xd;
of the individual ensemble partitions.  &#xd;
&#xd;
The quality of the consensus partitions highly depends on the ensemble&#xd;
generation mechanism and on the suitability of the consensus method for&#xd;
combining the generated ensemble.  Typically, consensus methods derive an&#xd;
ensemble representation that is used as the basis for extracting the consensus&#xd;
partition. Most ensemble representations circumvent the labeling problem.  On&#xd;
the other hand, voting-based methods establish direct parallels with consensus&#xd;
methods for supervised classifications, by seeking an optimal relabeling of the&#xd;
ensemble partitions and deriving an ensemble representation consisting of a&#xd;
central aggregated partition. An important element of the voting-based&#xd;
aggregation problem is the pairwise relabeling of an ensemble partition with&#xd;
respect to a representative partition of the ensemble, which is refered to here&#xd;
as the voting problem.  The voting problem is commonly formulated as a weighted&#xd;
bipartite matching problem.&#xd;
&#xd;
In this dissertation, a general theoretical framework for the voting problem as&#xd;
a multi-response regression problem is proposed. The problem is formulated as&#xd;
seeking to estimate the uncertainties associated with the assignments of the&#xd;
objects to the representative clusters, given their assignments to the clusters&#xd;
of an ensemble partition. A new voting scheme, referred to as cumulative voting,&#xd;
is derived as a special instance of the proposed regression formulation&#xd;
corresponding to fitting a linear model by least squares estimation. The&#xd;
proposed formulation reveals the close relationships between the underlying loss&#xd;
functions of the cumulative voting and bipartite matching schemes. A useful&#xd;
feature of the proposed framework is that it can be applied to model substantial&#xd;
variability between partitions, such as a variable number of clusters.  &#xd;
&#xd;
A general aggregation algorithm with variants corresponding to&#xd;
cumulative voting and bipartite matching is applied and a simulation-based&#xd;
analysis is presented to compare the suitability of each scheme to different&#xd;
ensemble generation mechanisms. The bipartite matching is found to be more&#xd;
suitable than cumulative voting for a particular generation model, whereby each&#xd;
ensemble partition is generated as a noisy permutation of an underlying&#xd;
labeling, according to a probability of error. For ensembles with a variable&#xd;
number of clusters, it is proposed that the aggregated partition be viewed as an&#xd;
estimated distributional representation of the ensemble, on the basis of which,&#xd;
a criterion may be defined to seek an optimally compressed consensus partition.&#xd;
&#xd;
The properties and features of the proposed cumulative voting scheme are&#xd;
studied.  In particular, the relationship between cumulative voting and the&#xd;
well-known co-association matrix is highlighted. Furthermore,  an adaptive&#xd;
aggregation algorithm that is suited for the cumulative voting scheme is&#xd;
proposed. The algorithm aims at selecting the initial reference partition and&#xd;
the aggregation sequence of the ensemble partitions the loss of mutual&#xd;
information associated with the aggregated partition is minimized.  In order to&#xd;
subsequently extract the final consensus partition, an efficient agglomerative&#xd;
algorithm is developed.  The algorithm merges the aggregated clusters such that&#xd;
the maximum amount of information is preserved. Furthermore, it allows the&#xd;
optimal number of consensus clusters to be estimated. &#xd;
&#xd;
An empirical study using several artificial and real-world datasets demonstrates&#xd;
that the proposed cumulative voting scheme leads to discovering substantially&#xd;
more accurate consensus partitions compared to bipartite matching, in the case&#xd;
of ensembles with a relatively large or a variable number of clusters. Compared&#xd;
to other recent consensus methods, the proposed method is found to be comparable&#xd;
with or better than the best performing methods. Moreover, accurate estimates of&#xd;
the true number of clusters are often achieved using cumulative voting, whereas&#xd;
consistently poor estimates are achieved based on bipartite matching. The&#xd;
empirical evidence demonstrates that the bipartite matching scheme is not&#xd;
suitable for these types of ensembles.</dim:field>
   <dim:field mdschema="dc" element="language" qualifier="iso" lang="en">en</dim:field>
   <dim:field mdschema="dc" element="publisher" lang="en">University of Waterloo</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en">Pattern Analysis</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en">Data Clustering</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en">Consensus Clustering</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en">Cluster Ensembles</dim:field>
   <dim:field mdschema="dc" element="subject" qualifier="program" lang="en">Electrical and Computer Engineering</dim:field>
   <dim:field mdschema="dc" element="title" lang="en">Voting-Based Consensus of Data Partitions</dim:field>
   <dim:field mdschema="dc" element="type" lang="en">Doctoral Thesis</dim:field>
   <dim:field mdschema="dc" element="pending" lang="en">false</dim:field>
   <dim:field mdschema="uws-etd" element="degree" qualifier="department" lang="en">Electrical and Computer Engineering</dim:field>
   <dim:field mdschema="uws-etd" element="degree" lang="en">Doctor of Philosophy</dim:field>
   <dim:field mdschema="uws" element="typeOfResource" lang="en">Text</dim:field>
   <dim:field mdschema="uws" element="peerReviewStatus" lang="en">Unreviewed</dim:field>
   <dim:field mdschema="uws" element="scholarLevel" lang="en">Graduate</dim:field>
   <dim:field mdschema="others" element="access-status">open.access</dim:field>
</dim:dim>
</metadata></record></GetRecord></OAI-PMH>