\relax \citation{weyeuker08,halstead77,mccabe76,chapman02,me04g,nagappan05,hall00,nikora03,conf/icse/NagappanB05a,khoshgoftaar01,conf/ictai/TangK04,journals/ese/KhoshgoftaarS03,me03a,me02e,me03k,me03q,porter90,tiang95,khoshgoftaar99,srinivasan95,me07b} \citation{me03q} \newlabel{@firstpg}{{}{1}} \@writefile{toc}{\contentsline {section}{\numberline {1}Introduction}{1}} \citation{me07b} \citation{lessmann09} \citation{milton08} \citation{me03q,me09b} \citation{me08d} \@writefile{toc}{\contentsline {section}{\numberline {2}Background}{2}} \newlabel{sec:back}{{2}{2}} \citation{leven95} \citation{lutz03} \citation{voas95} \citation{lowrey98} \citation{me99q,me00y} \citation{bradley98scaling} \citation{fenton99} \@writefile{toc}{\contentsline {subsection}{\numberline {2.1}Blind Spots}{3}} \@writefile{toc}{\contentsline {subsection}{\numberline {2.2}Lightweight Sampling}{3}} \@writefile{toc}{\contentsline {subsubsection}{\numberline {2.2.1}Data Mining}{3}} \citation{domingos97optimality} \citation{cohen95} \citation{quinlan92} \citation{briemann01,jiang08a} \@writefile{lof}{\contentsline {figure}{\numberline {1}{\ignorespaces Static code features.}}{4}} \newlabel{fig:features}{{1}{4}} \newlabel{sec:scf}{{2.2.2}{4}} \@writefile{toc}{\contentsline {subsubsection}{\numberline {2.2.2}Static Code Features}{4}} \citation{nikora04} \citation{fenton94} \citation{musa87,littlewood97} \citation{nach08} \citation{weyeuker08} \@writefile{toc}{\contentsline {subsection}{\numberline {2.3}Frequently Asked Questions}{5}} \@writefile{toc}{\contentsline {subsubsection}{\numberline {2.3.1}Why Binary Classifications?}{5}} \@writefile{toc}{\contentsline {subsubsection}{\numberline {2.3.2}Why Static Code Features?}{5}} \citation{nagappan05} \citation{me02f} \citation{rakitin01} \citation{fenton97} \citation{shepperd94} \@writefile{toc}{\contentsline {subsubsection}{\numberline {2.3.3}What Can be Learned from Static Code Features?}{6}} \citation{fenton97} \citation{me09b} \@writefile{lof}{\contentsline {figure}{\numberline {2}{\ignorespaces Tables of data.}}{7}} \newlabel{fig:data}{{2}{7}} \citation{me07e} \citation{me09b} \citation{me09b} \@writefile{lof}{\contentsline {figure}{\numberline {3}{\ignorespaces Performance measures}}{8}} \newlabel{fig:perform}{{3}{8}} \citation{shu02} \citation{fagan86} \citation{shull00a} \citation{fagan76} \citation{raffo05} \@writefile{lof}{\contentsline {figure}{\numberline {4}{\ignorespaces Results of round-robin and self experiments.}}{9}} \newlabel{fig:rrself}{{4}{9}} \@writefile{toc}{\contentsline {subsection}{\numberline {2.4}Summary}{9}} \newlabel{sec:summary}{{2.4}{9}} \citation{porter90} \citation{me07b} \citation{jiang08a} \citation{jiang08a} \citation{lessmann09} \citation{lessmann09} \citation{jiang08a} \citation{yang06} \citation{brieman96} \citation{FreSch97} \citation{cover67} \citation{quinlan92} \citation{cohen95r} \citation{breiman84} \citation{briemann01} \@writefile{lof}{\contentsline {figure}{\numberline {5}{\ignorespaces Box plot for AUC(pf,pd) seen with 9 learners.}}{10}} \newlabel{fig:logandnom}{{5}{10}} \@writefile{toc}{\contentsline {section}{\numberline {3}Ceiling Effects in Defect Predictors}{10}} \newlabel{sec:ceiling}{{3}{10}} \citation{lessmann09} \@writefile{lof}{\contentsline {figure}{\numberline {6}{\ignorespaces Range of AUC(pf,pd) ranks seen in 19 learners.}}{11}} \newlabel{fig:all}{{6}{11}} \@writefile{lof}{\contentsline {figure}{\numberline {7}{\ignorespaces $Pf$-vs-$pd$-vs-$effort$.}}{12}} \newlabel{fig:pdpf}{{7}{12}} \@writefile{toc}{\contentsline {section}{\numberline {4}Breaking Through the Ceiling}{12}} \newlabel{sec:b}{{4}{12}} \citation{arisholm06} \citation{jiang08a,jiang08b,me07b} \citation{cohen95} \citation{koru07} \@writefile{toc}{\contentsline {subsection}{\numberline {4.1}Experimental Set Up}{13}} \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.1.1}Operationalizing AUC(effort,pd)}{13}} \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.1.2}Upper and Lower Bounds on Performance}{13}} \@writefile{lof}{\contentsline {figure}{\numberline {8}{\ignorespaces $Effort$-vs-$PD$.}}{14}} \newlabel{fig:effort}{{8}{14}} \citation{quinlan92} \citation{elkan01} \citation{fawcett01} \citation{jiang08b,me08h} \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.1.3}Details}{15}} \@writefile{toc}{\contentsline {subsection}{\numberline {4.2}Initial results}{15}} \citation{milton08} \citation{Blake+Merz:1998} \citation{Blake+Merz:1998} \citation{Blake+Merz:1998} \citation{milton08} \citation{williams03} \citation{me02c,me03l} \@writefile{toc}{\contentsline {section}{\numberline {5}WHICH}{16}} \@writefile{toc}{\contentsline {subsection}{\numberline {5.1}Details}{16}} \@writefile{lof}{\contentsline {figure}{\numberline {9}{\ignorespaces Top-of-stack scores of the WHICH}}{17}} \newlabel{fig:picks}{{9}{17}} \newlabel{eq:pprime}{{1}{17}} \citation{jiang07,me00p,me02g,me02k,me03q,me04c,me05d,me07b,me07e} \citation{me08d} \newlabel{eq:weights}{{2}{18}} \@writefile{toc}{\contentsline {subsection}{\numberline {5.2}Results}{18}} \@writefile{toc}{\contentsline {subsubsection}{\numberline {5.2.1}Overall Results}{18}} \citation{me07b} \citation{jiang07,me00p,me02g,me02k,me03q,me04c,me05d,me07b,me07e} \@writefile{lof}{\contentsline {figure}{\numberline {10}{\ignorespaces Results from all data sets of Figure\nobreakspace {}2\hbox {}.}}{19}} \newlabel{fig:allr}{{10}{19}} \@writefile{toc}{\contentsline {subsubsection}{\numberline {5.2.2}Individual Results}{19}} \citation{lessmann09} \citation{me07b} \@writefile{lof}{\contentsline {figure}{\numberline {11}{\ignorespaces Four examples of pattern \#1}}{20}} \newlabel{fig:bands1a}{{11}{20}} \@writefile{lof}{\contentsline {figure}{\numberline {12}{\ignorespaces Three more examples of pattern \#1}}{21}} \newlabel{fig:bands1b}{{12}{21}} \@writefile{toc}{\contentsline {subsection}{\numberline {5.3}External Validity}{21}} \@writefile{lof}{\contentsline {figure}{\numberline {13}{\ignorespaces Two examples of pattern \#2.}}{22}} \newlabel{fig:bands2}{{13}{22}} \@writefile{lof}{\contentsline {figure}{\numberline {14}{\ignorespaces The only example of pattern \#3.}}{22}} \newlabel{fig:bands3}{{14}{22}} \citation{pugh90} \@writefile{toc}{\contentsline {section}{\numberline {6}Discussion}{23}} \@writefile{toc}{\contentsline {section}{\numberline {7}Conclusion}{23}} \bibstyle{acmtrans} \bibdata{refs} \bibcite{arisholm06}{\citeauthoryear {Arisholm and Briand}{Arisholm and Briand}{2006}} \bibcite{Blake+Merz:1998}{\citeauthoryear {Blake and Merz}{Blake and Merz}{1998}} \bibcite{bradley98scaling}{\citeauthoryear {Bradley, Fayyad, and Reina}{Bradley et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{1998}} \bibcite{breiman84}{\citeauthoryear {Breiman, Friedman, Olshen, and Stone}{Breiman et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{1984}} \bibcite{briemann01}{\citeauthoryear {Breimann}{Breimann}{2001}} \bibcite{brieman96}{\citeauthoryear {Brieman}{Brieman}{1996}} \bibcite{chapman02}{\citeauthoryear {Chapman and Solomon}{Chapman and Solomon}{2002}} \bibcite{cohen95}{\citeauthoryear {Cohen}{Cohen}{1995a}} \bibcite{cohen95r}{\citeauthoryear {Cohen}{Cohen}{1995b}} \bibcite{cover67}{\citeauthoryear {Cover and Hart}{Cover and Hart}{1967}} \bibcite{demsar06}{\citeauthoryear {Demsar}{Demsar}{2006}} \bibcite{dietterich97}{\citeauthoryear {Dietterich}{Dietterich}{1997}} \bibcite{domingos97optimality}{\citeauthoryear {Domingos and Pazzani}{Domingos and Pazzani}{1997}} \bibcite{elkan01}{\citeauthoryear {Elkan}{Elkan}{2001}} \bibcite{fagan76}{\citeauthoryear {Fagan}{Fagan}{1976}} \bibcite{fagan86}{\citeauthoryear {Fagan}{Fagan}{1986}} \bibcite{fawcett01}{\citeauthoryear {Fawcett}{Fawcett}{2001}} \bibcite{fenton94}{\citeauthoryear {Fenton, Pfleeger, and Glass}{Fenton et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{1994}} \bibcite{fenton99}{\citeauthoryear {Fenton and Neil}{Fenton and Neil}{1999}} \bibcite{fenton96}{\citeauthoryear {Fenton and Pfleeger}{Fenton and Pfleeger}{1995}} \bibcite{fenton97}{\citeauthoryear {Fenton and Pfleeger}{Fenton and Pfleeger}{1997}} \bibcite{FreSch97}{\citeauthoryear {Freund and Schapire}{Freund and Schapire}{1997}} \bibcite{hall00}{\citeauthoryear {Hall and Munson}{Hall and Munson}{2000}} \bibcite{halstead77}{\citeauthoryear {Halstead}{Halstead}{1977}} \bibcite{huang05}{\citeauthoryear {Huang and Ling}{Huang and Ling}{2005}} \bibcite{jiang08b}{\citeauthoryear {Jiang, Cukic, and Ma}{Jiang et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2008}} \bibcite{jiang07}{\citeauthoryear {Jiang, Cukic, and Menzies}{Jiang et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2007}} \bibcite{jiang08a}{\citeauthoryear {Jiang, Cukic, and Menzies}{Jiang et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2008}} \bibcite{khoshgoftaar01}{\citeauthoryear {Khoshgoftaar}{Khoshgoftaar}{2001}} \bibcite{khoshgoftaar99}{\citeauthoryear {Khoshgoftaar and Allen}{Khoshgoftaar and Allen}{2001}} \bibcite{journals/ese/KhoshgoftaarS03}{\citeauthoryear {Khoshgoftaar and Seliya}{Khoshgoftaar and Seliya}{2003}} \bibcite{koru07}{\citeauthoryear {Koru, Zhang, and Liu}{Koru et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2007}} \bibcite{lessmann09}{\citeauthoryear {Lessmann, Baesens, Mues, and Pietsch}{Lessmann et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2008}} \bibcite{leven95}{\citeauthoryear {Leveson}{Leveson}{1995}} \bibcite{littlewood97}{\citeauthoryear {Littlewood and Wright}{Littlewood and Wright}{1997}} \bibcite{lowrey98}{\citeauthoryear {Lowry, Boyd, and Kulkarni}{Lowry et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{1998}} \bibcite{lutz03}{\citeauthoryear {Lutz and Mikulski}{Lutz and Mikulski}{2003}} \bibcite{mccabe76}{\citeauthoryear {McCabe}{McCabe}{1976}} \bibcite{me00p}{\citeauthoryear {Menzies}{Menzies}{2001}} \bibcite{me02c}{\citeauthoryear {Menzies, Chiang, Feather, Hu, and Kiper}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2003}} \bibcite{me99q}{\citeauthoryear {Menzies and Cukic}{Menzies and Cukic}{2000}} \bibcite{me00y}{\citeauthoryear {Menzies and Cukic}{Menzies and Cukic}{2002}} \bibcite{me07e}{\citeauthoryear {Menzies, Dekhtyar, Distefano, and Greenwald}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2007}} \bibcite{me04g}{\citeauthoryear {Menzies, DiStefano, Orrego, and Chapman}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2004}} \bibcite{me02e}{\citeauthoryear {Menzies, DiStefeno, Chapman, and Mcgill}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2002}} \bibcite{me07b}{\citeauthoryear {Menzies, Greenwald, and Frank}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2007}} \bibcite{me02g}{\citeauthoryear {Menzies, Lutz, and Mikulski}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2003}} \bibcite{me05d}{\citeauthoryear {Menzies, Port, Chen, Hihn, and Stukes}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2005}} \bibcite{me02f}{\citeauthoryear {Menzies, Raffo, on\nobreakspace {}Setamanit, Hu, and Tootoonian}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2002}} \bibcite{me03l}{\citeauthoryear {Menzies and Singh}{Menzies and Singh}{2003}} \bibcite{me02k}{\citeauthoryear {Menzies and Stefano}{Menzies and Stefano}{2003a}} \bibcite{me03a}{\citeauthoryear {Menzies, Stefano, Ammar, McGill, Callis, Chapman, and J}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2003}} \bibcite{me03k}{\citeauthoryear {Menzies, Stefano, and Chapman}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2003}} \bibcite{me03q}{\citeauthoryear {Menzies and Stefano}{Menzies and Stefano}{2003b}} \bibcite{me04c}{\citeauthoryear {Menzies, Stefano, Cunanan, and Chapman}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2004}} \bibcite{me08d}{\citeauthoryear {Menzies, Turhan, Bener, Gay, Cukic, and Jiang}{Menzies et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2008}} \bibcite{milton08}{\citeauthoryear {Milton}{Milton}{2008}} \bibcite{musa87}{\citeauthoryear {Musa, Iannino, and Okumoto}{Musa et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{1987}} \bibcite{nagappan05}{\citeauthoryear {Nagappan and Ball}{Nagappan and Ball}{2005a}} \bibcite{conf/icse/NagappanB05a}{\citeauthoryear {Nagappan and Ball}{Nagappan and Ball}{2005b}} \bibcite{nach08}{\citeauthoryear {Nagappan, Murphy, and V}{Nagappan et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2008}} \bibcite{nikora04}{\citeauthoryear {Nikora}{Nikora}{2004}} \bibcite{nikora03}{\citeauthoryear {Nikora and Munson}{Nikora and Munson}{2003}} \bibcite{porter90}{\citeauthoryear {Porter and Selby}{Porter and Selby}{1990}} \bibcite{pugh90}{\citeauthoryear {Pugh}{Pugh}{1990}} \bibcite{quinlan92}{\citeauthoryear {Quinlan}{Quinlan}{1992}} \bibcite{raffo05}{\citeauthoryear {Raffo}{Raffo}{2005}} \bibcite{rakitin01}{\citeauthoryear {Rakitin}{Rakitin}{2001}} \bibcite{shepperd94}{\citeauthoryear {Shepperd and Ince}{Shepperd and Ince}{1994}} \bibcite{shu02}{\citeauthoryear {Shull, ad\nobreakspace {}B.\nobreakspace {}Boehm, Brown, Costa, Lindvall, Port, Rus, Tesoriero, and Zelkowitz}{Shull et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2002}} \bibcite{shull00a}{\citeauthoryear {Shull, Rus, and Basili}{Shull et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2000}} \bibcite{srinivasan95}{\citeauthoryear {Srinivasan and Fisher}{Srinivasan and Fisher}{1995}} \bibcite{conf/ictai/TangK04}{\citeauthoryear {Tang and Khoshgoftaar}{Tang and Khoshgoftaar}{2004}} \bibcite{tiang95}{\citeauthoryear {Tian and Zelkowitz}{Tian and Zelkowitz}{1995}} \bibcite{me09b}{\citeauthoryear {Turhan, Menzies, Bener, and Distefano}{Turhan et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2009}} \bibcite{voas95}{\citeauthoryear {Voas and Miller}{Voas and Miller}{1995}} \bibcite{weyeuker08}{\citeauthoryear {Weyuker, Ostrand, and Bell}{Weyuker et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2008}} \bibcite{williams03}{\citeauthoryear {Williams, Gomes, and Selman}{Williams et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2003}} \bibcite{witten05}{\citeauthoryear {Witten and Frank}{Witten and Frank}{2005}} \bibcite{me08h}{\citeauthoryear {Y.\nobreakspace {}Jiang and Menzies}{Y.\nobreakspace {}Jiang and Menzies}{2008}} \bibcite{yang06}{\citeauthoryear {Yang, Webb, Cerquides, Korb, Boughton, and Ting}{Yang et\nobreakspace {}al\unhbox \voidb@x \hbox {.}}{2006}} \citation{witten05} \citation{witten05} \citation{domingos97optimality} \citation{domingos97optimality} \citation{quinlan92} \citation{witten05} \citation{quinlan92} \citation{cohen95r} \citation{dietterich97} \citation{halstead77} \citation{mccabe76} \citation{fenton96} \citation{fenton96} \citation{fenton96} \citation{demsar06} \citation{demsar06,huang05} \citation{demsar06} \newlabel{@lastpg}{{7}{30}}