\relax \citation{clayton98wcre} \citation{wiggerts97wcre2} \@writefile{toc}{\contentsline {section}{\numberline {1}\hskip -1em.\penalty \@M \ Introduction}{1}} \citation{lakhotia97JSS} \citation{lakhotia97JSS} \citation{kunz95TSE} \citation{patel92icse} \citation{canfora94icsm} \citation{muller93SMRP} \citation{tzerpos97wcre} \citation{anquetil95oopsla} \citation{mancoridis96icsm} \citation{tzerpos97wcre} \citation{anquetil99JSMRP} \@writefile{toc}{\contentsline {section}{\numberline {2}Issues in Clustering}{2}} \newlabel{sec:clust}{{2}{2}} \@writefile{toc}{\contentsline {subsection}{\numberline {2.1}Entities to Cluster and their Descriptions}{2}} \newlabel{sec:clust-desc}{{2.1}{2}} \@writefile{lof}{\contentsline {figure}{\numberline {1}{\ignorespaces Some files and possible clusters based on their inclusion.}}{2}} \newlabel{fig:includes}{{1}{2}} \citation{sneath73} \citation{sneath73} \citation{sneath73} \@writefile{toc}{\contentsline {subsection}{\numberline {2.2}Similarity Metrics}{3}} \newlabel{sec:clust-dist}{{2.2}{3}} \citation{lakhotia97JSS} \citation{wiggerts97wcre2} \citation{wiggerts97wcre2} \@writefile{toc}{\contentsline {subsection}{\numberline {2.3}Clustering Algorithms}{4}} \newlabel{sec:clust-algo}{{2.3}{4}} \@writefile{lof}{\contentsline {figure}{\numberline {2}{\ignorespaces Distance of a new cluster (\emph {C4}) to other clusters (\emph {C3}) in agglomerative hierarchical clustering algorithms.}}{4}} \newlabel{fig:algo-dist}{{2}{4}} \@writefile{lof}{\contentsline {figure}{\numberline {3}{\ignorespaces Influence of the clustering algorithm on the clusters (from \cite {wiggerts97wcre2}). Left: Single linkage. Right: Complete linkage.}}{4}} \newlabel{fig:sing-comp}{{3}{4}} \@writefile{lot}{\contentsline {table}{\numberline {1}{\ignorespaces Summary of the three clustering parameters studied. ``Zero'' identifies the similarity metrics which do consider null-dimensions.}}{5}} \newlabel{tab:sum-params}{{1}{5}} \@writefile{toc}{\contentsline {subsection}{\numberline {2.4}System Partitions}{5}} \newlabel{sec:clust-part}{{2.4}{5}} \@writefile{lof}{\contentsline {figure}{\numberline {4}{\ignorespaces A hierarchy of clusters and how to cut it to get a partition of the data set.}}{5}} \newlabel{fig:clust-cut}{{4}{5}} \@writefile{toc}{\contentsline {section}{\numberline {3}Quality Criteria}{5}} \newlabel{sec:quality}{{3}{5}} \citation{salton83} \citation{sommerville95} \citation{kunz95TSE} \citation{patel92icse} \@writefile{toc}{\contentsline {subsection}{\numberline {3.1}Expert Criterion}{6}} \newlabel{sec:qual-expert}{{3.1}{6}} \@writefile{toc}{\contentsline {subsection}{\numberline {3.2}Design Criterion}{6}} \newlabel{sec:qual-design}{{3.2}{6}} \citation{mancoridis98iwpc} \@writefile{lof}{\contentsline {figure}{\numberline {5}{\ignorespaces Size criterion for two algorithms. The algorithm with more intermediate clusters (complete linkage on the left) is better. System: gcc. Feature: File inclusions. Similarity metric: Jaccard coefficient.}}{7}} \newlabel{fig:qual-size}{{5}{7}} \@writefile{toc}{\contentsline {subsection}{\numberline {3.3}Size Criterion}{7}} \newlabel{sec:qual-size}{{3.3}{7}} \citation{mosaic} \citation{gcc} \@writefile{lof}{\contentsline {figure}{\numberline {6}{\ignorespaces Comparison of four algorithms using the design criterion. System: Linux. Feature: file inclusions. Similarity metric: correlation.}}{8}} \newlabel{fig:algos-design}{{6}{8}} \@writefile{lof}{\contentsline {figure}{\numberline {7}{\ignorespaces Comparison of four algorithms using the expert criterion. System: Linux. Feature: file inclusions. Similarity metric: correlation.}}{8}} \newlabel{fig:algos-expert}{{7}{8}} \@writefile{toc}{\contentsline {section}{\numberline {4}Results}{8}} \newlabel{sec:result}{{4}{8}} \@writefile{toc}{\contentsline {subsection}{\numberline {4.1}Clustering Algorithms}{8}} \newlabel{sec:res-algo}{{4.1}{8}} \citation{wiggerts97wcre2} \@writefile{lof}{\contentsline {figure}{\numberline {8}{\ignorespaces Comparison of three similarity metrics using the size criterion. System: Mosaic; Feature: type references; Algorithm: Complete linkage.}}{9}} \newlabel{fig:dist-size}{{8}{9}} \@writefile{toc}{\contentsline {subsection}{\numberline {4.2}Similarity Metrics}{9}} \newlabel{sec:res-dist}{{4.2}{9}} \@writefile{toc}{\contentsline {subsection}{\numberline {4.3}Entity Descriptions}{9}} \newlabel{sec:res-desc}{{4.3}{9}} \@writefile{toc}{\contentsline {section}{\numberline {5}Related Work}{9}} \newlabel{sec:relat}{{5}{9}} \citation{lakhotia97JSS} \citation{armstrong98wcre} \citation{storey96wcre} \citation{girard97wcre} \citation{deursen99icse} \citation{lindig97icse} \citation{siff97icsm} \citation{deursen99icse} \citation{anquetil94} \citation{clayton98wcre} \@writefile{lof}{\contentsline {figure}{\numberline {9}{\ignorespaces Comparison of four different features using the design criterion. Earliest cuts (height $<0.5$) were ignored to improve readability. System: Mosaic. Algorithm: Complete linkage. Similarity metric: Jaccard coefficient.}}{10}} \newlabel{fig:desc-design}{{9}{10}} \@writefile{lof}{\contentsline {figure}{\numberline {10}{\ignorespaces Comparison of four different features using the expert criterion. Earliest cuts (height $<0.5$) were ignored to improve readability. System: Mosaic. Algorithm: Complete linkage. Similarity metric: Jaccard coefficient.}}{10}} \newlabel{fig:desc-expert}{{10}{10}} \citation{wiggerts97wcre2} \citation{deursen99icse} \bibstyle{latex8} \bibdata{kbre,reference,doc,a-lire} \bibcite{anquetil99JSMRP}{1} \bibcite{anquetil94}{2} \bibcite{anquetil95oopsla}{3} \@writefile{toc}{\contentsline {section}{\numberline {6}Conclusion}{11}} \bibcite{deursen99icse}{4} \bibcite{armstrong98wcre}{5} \bibcite{canfora94icsm}{6} \bibcite{clayton98wcre}{7} \bibcite{gcc}{8} \bibcite{girard97wcre}{9} \bibcite{kunz95TSE}{10} \bibcite{lakhotia97JSS}{11} \bibcite{lindig97icse}{12} \bibcite{mancoridis96icsm}{13} \bibcite{mancoridis98iwpc}{14} \bibcite{mosaic}{15} \bibcite{muller93SMRP}{16} \bibcite{patel92icse}{17} \bibcite{salton83}{18} \bibcite{siff97icsm}{19} \bibcite{sneath73}{20} \bibcite{sommerville95}{21} \bibcite{storey96wcre}{22} \bibcite{tzerpos97wcre}{23} \bibcite{wiggerts97wcre2}{24}