电脑显示stop:c000021aunknownharderrorunknokwnharderr_百度知道
电脑显示stop:c000021aunknownharderrorunknokwnharderr
谢谢，我在卸载软件
卸载软件蓝屏，重启电脑一般就没事了，不会总蓝屏吧？有问题您就直接追问。您的电脑蓝屏的时候，您在电脑干什么呢，能说说吗？我会跟据您说的较为准确的回答您。 蓝屏代码或事件查看器里面的内容普通人是看不懂的，请将你在蓝屏前电脑的表现，和你操作说的详细些（我跟据你提供的信息重新回答你）。 一般蓝屏是自己不正确操作引起的，记住容易引起蓝屏的操作不做。电脑不要满负荷操作，就是在玩游戏、看视频时、下载时、看网页的同时在干别的操作最容易死机、蓝屏，因此在玩游戏、看视频、下载时、看网页时不要在操作别的东西了。 不管您在干什么，只要一有卡的迹象时就赶紧停止手头的操作退出在试，这样就不会蓝屏，如果还是这样就不要玩了或是不要看了。 硬件方面：如果内存小请加内存条，硬盘是否有坏道，硬件是否不兼容或是故障，在用鲁大师测试一下CPU等硬件的温度是否高。 90%的蓝屏是自己操作使用电脑不当引起的，卡是蓝屏发生的根源，容易引起卡的操作不做，就会避免蓝屏或少发生蓝屏。
其他类似问题
为您推荐：
电脑显示的相关知识
等待您来回答
下载知道APP
随时随地咨询
出门在外也不愁您所在位置： &
&nbsp&&nbsp&nbsp&&nbsp
基于OPC+DA技术的客户应用软件的设计及实现.pdf58页
本文档一共被下载：
次 ,您可免费全文在线阅读后下载本文档
文档加载中...广告还剩秒
需要金币：200 &&
你可能关注的文档：
··········
··········
forProcessCon仃0l，即过程控
OPC全称是OLE ObjectLinking锄dEmbedding
分布式组件对象模型 技术所要求的功能，制定的一个开放和互用式的工控行业的软件
接口标准，在当今过程控制领域，是一种非常流行的数据交换技术，是实现控制系统现
场设备级与过程管理级信息交互、实现控制系统开放性的中间件技术。
论文首先介绍了OPC技术产生的背景，以及它的主要特点与发展状况，深入剖析
了其核心COM／DCOM技术，然后重点分析了OPC数据访问规范 OPCDA 及OPC客
户端与服务器的通信机制。本文基于OPC数据访问规范及COM／DCOM理论，主要做
了一下几个方面的工作：
1 提出用OPC技术解决工艺装置与调度部门之间的实时数据传递问题。开发了
OPC实时数据客户端软件，解决了工厂问题，达到了预期效果。
技术问题。
3 针对OPC数据访问规范讨论了如何进行源码级的开发及实现，在充分考虑OPC
DA2．O版与1．O旧版的不同之处，讨论了兼容新旧版本的客户端软件的实现方法，实现
了对本地或远方计算机的OPC服务器进行数据采集和管理维护功能。
4 利用EXCEL自动化接口技术，讨论了实现实时报表功能。
总之，课题的研究和实践致力于对理解OPC内部机制和开发OPC客户端具体应用
提供良好的参考，同时展现了COM／DCOM中定制接口和自动化接口技术的使用，促进
基于0PC中间件技术来构建现代信息控制系统的发展。
关键词：OPC；COM／DCOM：数据访问规范；OPC客户端；报表
Con钾ol，l
正在加载中，请稍后...%!PS-Adobe-2.0
%%Creator: dvipsk 5.58f Copyright
Radical Eye Software
%%Title: final.dvi
%%Pages: 9
%%PageOrder: Ascend
%%BoundingBox: 0 0 612 792
%%EndComments
%DVIPSCommandLine: dvips final
%DVIPSParameters: dpi=300, compressed, comments removed
%DVIPSSource:
TeX output :1117
%%BeginProcSet: texc.pro
/TeXDict 250 dict def TeXDict begin /N{def}def /B{bind def}N /S{exch}N
/X{S N}B /TR{translate}N /isls false N /vsize 11 72 mul N /hsize 8.5 72
mul N /landplus90{false}def /@rigin{isls{[0 landplus90{1 -1}{-1 1}
ifelse 0 0 0]concat}if 72 Resolution div 72 VResolution div neg scale
isls{landplus90{VResolution 72 div vsize mul 0 exch}{Resolution -72 div
hsize mul 0}ifelse TR}if Resolution VResolution vsize -72 div 1 add mul
TR[matrix currentmatrix{dup dup round sub abs 0.00001 lt{round}if}
forall round exch round exch]setmatrix}N /@landscape{/isls true N}B
/@manualfeed{statusdict /manualfeed true put}B /@copies{/#copies X}B
/FMat[1 0 0 -1 0 0]N /FBB[0 0 0 0]N /nn 0 N /IE 0 N /ctr 0 N /df-tail{
/nn 8 dict N nn begin /FontType 3 N /FontMatrix fntrx N /FontBBox FBB N
string /base X array /BitMaps X /BuildChar{CharBuilder}N /Encoding IE N
end dup{/foo setfont}2 array copy cvx N load 0 nn put /ctr 0 N[}B /df{
/sf 1 N /fntrx FMat N df-tail}B /dfs{div /sf X /fntrx[sf 0 0 sf neg 0 0]
N df-tail}B /E{pop nn dup definefont setfont}B /ch-width{ch-data dup
length 5 sub get}B /ch-height{ch-data dup length 4 sub get}B /ch-xoff{
128 ch-data dup length 3 sub get sub}B /ch-yoff{ch-data dup length 2 sub
get 127 sub}B /ch-dx{ch-data dup length 1 sub get}B /ch-image{ch-data
dup type /stringtype ne{ctr get /ctr ctr 1 add N}if}B /id 0 N /rw 0 N
/rc 0 N /gp 0 N /cp 0 N /G 0 N /sf 0 N /CharBuilder{save 3 1 roll S dup
/base get 2 index get S /BitMaps get S get /ch-data X pop /ctr 0 N ch-dx
0 ch-xoff ch-yoff ch-height sub ch-xoff ch-width add ch-yoff
setcachedevice ch-width ch-height true[1 0 0 -1 -.1 ch-xoff sub ch-yoff
.1 sub]/id ch-image N /rw ch-width 7 add 8 idiv string N /rc 0 N /gp 0 N
/cp 0 N{rc 0 ne{rc 1 sub /rc X rw}{G}ifelse}imagemask restore}B /G{{id
gp get /gp gp 1 add N dup 18 mod S 18 idiv pl S get exec}loop}B /adv{cp
add /cp X}B /chg{rw cp id gp 4 index getinterval putinterval dup gp add
/gp X adv}B /nd{/cp 0 N rw exit}B /lsh{rw cp 2 copy get dup 0 eq{pop 1}{
dup 255 eq{pop 254}{dup dup add 255 and S 1 and or}ifelse}ifelse put 1
adv}B /rsh{rw cp 2 copy get dup 0 eq{pop 128}{dup 255 eq{pop 127}{dup 2
idiv S 128 and or}ifelse}ifelse put 1 adv}B /clr{rw cp 2 index string
putinterval adv}B /set{rw cp fillstr 0 4 index getinterval putinterval
adv}B /fillstr 18 string 0 1 17{2 copy 255 put pop}for N /pl[{adv 1 chg}
{adv 1 chg nd}{1 add chg}{1 add chg nd}{adv lsh}{adv lsh nd}{adv rsh}{
adv rsh nd}{1 add adv}{/rc X nd}{1 add set}{1 add clr}{adv 2 chg}{adv 2
chg nd}{pop nd}]dup{bind pop}forall N /D{/cc X dup type /stringtype ne{]
}if nn /base get cc ctr put nn /BitMaps get S ctr S sf 1 ne{dup dup
length 1 sub dup 2 index S get sf div put}if put /ctr ctr 1 add N}B /I{
cc 1 add D}B /bop{userdict /bop-hook known{bop-hook}if /SI save N @rigin
0 0 moveto /V matrix currentmatrix dup 1 get dup mul exch 0 get dup mul
add .99 lt{/QV}{/RV}ifelse load def pop pop}N /eop{SI restore userdict
/eop-hook known{eop-hook}if showpage}N /@start{userdict /start-hook
known{start-hook}if pop /VResolution X /Resolution X 1000 div /DVImag X
/IE 256 array N 0 1 255{IE S 1 string dup 0 3 index put cvn put}for
65781.76 div /vsize X 65781.76 div /hsize X}N /p{show}N /RMat[1 0 0 -1 0
0]N /BDot 260 string N /rulex 0 N /ruley 0 N /v{/ruley X /rulex X V}B /V
{}B /RV statusdict begin /product where{pop product dup length 7 ge{0 7
getinterval dup(Display)eq exch 0 4 getinterval(NeXT)eq or}{pop false}
ifelse}{false}ifelse end{{gsave TR -.1 .1 TR 1 1 scale rulex ruley false
RMat{BDot}imagemask grestore}}{{gsave TR -.1 .1 TR rulex ruley scale 1 1
false RMat{BDot}imagemask grestore}}ifelse B /QV{gsave newpath transform
round exch round exch itransform moveto rulex 0 rlineto 0 ruley neg
rlineto rulex neg 0 rlineto fill grestore}B /a{moveto}B /delta 0 N /tail
{dup /delta X 0 rmoveto}B /M{S p delta add tail}B /b{S p tail}B /c{-4 M}
B /d{-3 M}B /e{-2 M}B /f{-1 M}B /g{0 M}B /h{1 M}B /i{2 M}B /j{3 M}B /k{
4 M}B /w{0 rmoveto}B /l{p -4 w}B /m{p -3 w}B /n{p -2 w}B /o{p -1 w}B /q{
p 1 w}B /r{p 2 w}B /s{p 3 w}B /t{p 4 w}B /x{0 S rmoveto}B /y{3 2 roll p
a}B /bos{/SS save N}B /eos{SS restore}B end
%%EndProcSet
%%BeginProcSet: special.pro
TeXDict begin /SDict 200 dict N SDict begin /@SpecialDefaults{/hs 612 N
/vs 792 N /ho 0 N /vo 0 N /hsc 1 N /vsc 1 N /ang 0 N /CLIP 0 N /rwiSeen
false N /rhiSeen false N /letter{}N /note{}N /a4{}N /legal{}N}B
/@scaleunit 100 N /@hscale{@scaleunit div /hsc X}B /@vscale{@scaleunit
div /vsc X}B /@hsize{/hs X /CLIP 1 N}B /@vsize{/vs X /CLIP 1 N}B /@clip{
/CLIP 2 N}B /@hoffset{/ho X}B /@voffset{/vo X}B /@angle{/ang X}B /@rwi{
10 div /rwi X /rwiSeen true N}B /@rhi{10 div /rhi X /rhiSeen true N}B
/@llx{/llx X}B /@lly{/lly X}B /@urx{/urx X}B /@ury{/ury X}B /magscale
true def end /@MacSetUp{userdict /md known{userdict /md get type
/dicttype eq{userdict begin md length 10 add md maxlength ge{/md md dup
length 20 add dict copy def}if end md begin /letter{}N /note{}N /legal{}
N /od{txpose 1 0 mtx defaultmatrix dtransform S atan/pa X newpath
clippath mark{transform{itransform moveto}}{transform{itransform lineto}
}{6 -2 roll transform 6 -2 roll transform 6 -2 roll transform{
itransform 6 2 roll itransform 6 2 roll itransform 6 2 roll curveto}}{{
closepath}}pathforall newpath counttomark array astore /gc xdf pop ct 39
0 put 10 fz 0 fs 2 F/|______Courier fnt invertflag{PaintBlack}if}N
/txpose{pxs pys scale ppr aload pop por{noflips{pop S neg S TR pop 1 -1
scale}if xflip yflip and{pop S neg S TR 180 rotate 1 -1 scale ppr 3 get
ppr 1 get neg sub neg ppr 2 get ppr 0 get neg sub neg TR}if xflip yflip
not and{pop S neg S TR pop 180 rotate ppr 3 get ppr 1 get neg sub neg 0
TR}if yflip xflip not and{ppr 1 get neg ppr 0 get neg TR}if}{noflips{TR
pop pop 270 rotate 1 -1 scale}if xflip yflip and{TR pop pop 90 rotate 1
-1 scale ppr 3 get ppr 1 get neg sub neg ppr 2 get ppr 0 get neg sub neg
TR}if xflip yflip not and{TR pop pop 90 rotate ppr 3 get ppr 1 get neg
sub neg 0 TR}if yflip xflip not and{TR pop pop 270 rotate ppr 2 get ppr
0 get neg sub neg 0 S TR}if}ifelse scaleby96{ppr aload pop 4 -1 roll add
2 div 3 1 roll add 2 div 2 copy TR .96 dup scale neg S neg S TR}if}N /cp
{pop pop showpage pm restore}N end}if}if}N /normalscale{Resolution 72
div VResolution 72 div neg scale magscale{DVImag dup scale}if 0 setgray}
N /psfts{S 65781.76 div N}N /startTexFig{/psf$SavedState save N userdict
maxlength dict begin /magscale true def normalscale currentpoint TR
/psf$ury psfts /psf$urx psfts /psf$lly psfts /psf$llx psfts /psf$y psfts
/psf$x psfts currentpoint /psf$cy X /psf$cx X /psf$sx psf$x psf$urx
psf$llx sub div N /psf$sy psf$y psf$ury psf$lly sub div N psf$sx psf$sy
scale psf$cx psf$sx div psf$llx sub psf$cy psf$sy div psf$ury sub TR
/showpage{}N /erasepage{}N /copypage{}N /p 3 def @MacSetUp}N /doclip{
psf$llx psf$lly psf$urx psf$ury currentpoint 6 2 roll newpath 4 copy 4 2
roll moveto 6 -1 roll S lineto S lineto S lineto closepath clip newpath
moveto}N /endTexFig{end psf$SavedState restore}N /@beginspecial{SDict
begin /SpecialSave save N gsave normalscale currentpoint TR
@SpecialDefaults count /ocount X /dcount countdictstack N}N /@setspecial
{CLIP 1 eq{newpath 0 0 moveto hs 0 rlineto 0 vs rlineto hs neg 0 rlineto
closepath clip}if ho vo TR hsc vsc scale ang rotate rwiSeen{rwi urx llx
sub div rhiSeen{rhi ury lly sub div}{dup}ifelse scale llx neg lly neg TR
}{rhiSeen{rhi ury lly sub div dup scale llx neg lly neg TR}if}ifelse
CLIP 2 eq{newpath llx lly moveto urx lly lineto urx ury lineto llx ury
lineto closepath clip}if /showpage{}N /erasepage{}N /copypage{}N newpath
}N /@endspecial{count ocount sub{pop}repeat countdictstack dcount sub{
end}repeat grestore SpecialSave restore end}N /@defspecial{SDict begin}
N /@fedspecial{end}B /li{lineto}B /rl{rlineto}B /rc{rcurveto}B /np{
/SaveX currentpoint /SaveY X N 1 setlinecap newpath}N /st{stroke SaveX
SaveY moveto}N /fil{fill SaveX SaveY moveto}N /ellipse{/endangle X
/startangle X /yrad X /xrad X /savematrix matrix currentmatrix N TR xrad
yrad scale 0 0 1 startangle endangle arc savematrix setmatrix}N end
%%EndProcSet
TeXDict begin
300 300 (final.dvi)
@start /Fa 4 63 df<F0EA01E0EA03C0EAA121E123E123C12
7CAB3A50D D<12F8B3A23C123E121E121F7EEA07
80EA03C0EA01E0EA00FD D<137CB3A613F8A313F
A5A12F012C012FE7EE00
13F8AE4D D<12F8AE050E D E
/Fb 1 50 df49 D E /Fc
3 51 df43 D49 DI E /Fd 29 122
df<EBFE7CA6B512FAE38FF8FF0171A
D40 DI44 D46 D58 D67 D72 D84 D87 D97 DIIII
IIII108 D110 DII<EAFC78EA1D9C
EA1E1CABEAFF800E4 DIIII121
D E /Fe 24 123 df13
DI97 D99 DII<EB38C013C5EA0E
A30EA3EA0C1EEA047CEA039CEA001CA25B1260EAF0301370EAE0C0007FC7
FC121A D105 D107 D109 DI<C860
E0EA11CC0EBEA0E86
FCA25AA1A D114 DII118 DIIII
E /Ff 7 123 df13 D59
D84 D118 DI121 DI
E /Fg 7 104 df0 D2 D15 D
20 DI102 DI E
/Fh 49 123 df
38 D45 DII49
IIII72 DII76 DIIIIIIII97 DIIIIIIII107
IIIII E /Fi 31 117 df49
DIII65 D<D903FEFFF81
01F31FC1127EA2
00FE91C7FCA8127EEDECB6C6C13386CB4
13F090381FFFC0D903FEC7FC2 DIII73 D75
DIIIII<B512FEECFFCEC01F86E7E157E157FAA5A
EC07F090B512C05DE7E6E7E6E7EA81606EC00FEEDFF0CBED0FF0
2 DIII<B53A0FFFF00F
FEAC712EC0A26C6CA26D14C0A26D49
6C5A6AF00CA3F0C301FC38011F15
FEFEEBFFE06D486D5AA26D486D5AA36D486DC7FCA36D48130EA2
101 DI110 D114 DII E /Fj 82 124 df11
34 D37 D39
DIIIIIIIIIIIIIIII61 D64 DIIIIIIIIIIIIIIIIIIII
D97 DIIIIIIIIIIIIIIIIIIIIIIIIII E /Fk 45 122 df<EAFFF8A40D DI49
DIII58 D65 D67 DIIIIII76
DIIII82 DIII87 DI97 DIIII<0E1F38EA3C0F
00EA3C0FEA1E1CEA33F00030C7FCA3EA3FFF14C06C13E014F
F81E03C0B7F
DII107 DIIII114
DIII<39FF8FF8FEAAF8E03907CC78
C0A2EBF83D143FAEEBE00EA21F127F
D<38FFC1FCAA803E300A213F7EA01F613FE6C5A
AA25B1270EAF8E05BEAFFC123E161A7F
D E /Fl 19 119 df65
D99 DII103 DII110 DII114 DIIII
%%EndProlog
%%BeginSetup
%%Feature: *Resolution 300dpi
TeXDict begin
%%EndSetup
%%Page: 1 1
1 0 bop 15 116 2025 5 v 167 233 a Fl(An)23 b(Inductiv)n(e)i(Learning)e
(Approac)n(h)g(to)f(Prognostic)h(Prediction)p 15 324
V 221 463 a Fk(W.)16 b(Nic)o(k)g(Street)114 513 y Fj(Departmen)o(ts)d
(of)g(Surgery)i(and)199 563 y(Computer)e(Sciences)156
613 y(Univ)o(ersit)o(y)h(of)f(Wisconsin)143 663 y(1210)g(W)m(est)h(Da)o
(yton)f(Street)194 712 y(Madison,)g(WI)h(2
y(street@cs.wisc.edu)833 463 y Fk(O.)i(L.)g(Mangasarian)716
513 y Fj(Departmen)o(t)e(of)f(Computer)g(Sciences)812
563 y(Univ)o(ersit)o(y)h(of)f(Wisconsin)799 613 y(1210)g(W)m(est)h(Da)o
(yton)f(Street)850 663 y(Madison,)g(WI)h(2
y(olvi@cs.wisc.edu) y Fk(W.)j(H.)f(W)l(olb)q(erg)1475
513 y Fj(Departmen)o(t)d(of)g(Surgery) y(Univ)o(ersit)o(y)h(of)
f(Wisconsin) y(600)h(Highland)e(Av)o(en)o(ue)1507
663 y(Madison,)h(WI)g( 712 y(w)o(olb)q(erg@eagle.surgery)m
(.wisc.edu)395 866 y Fi(Abstract)98 994 y Fj(This)21
b(pap)q(er)g(in)o(tro)q(duces)h(the)f(Recurrence)i(Sur-)98
1044 y(face)h(Appro)o(ximation,)g(an)g(inductiv)o(e)g(learning)98
1094 y(metho)q(d)c(based)i(on)f(linear)g(programming)d(that)98
1143 y(predicts)34 b(recurrence)i(times)c(using)h(censored)98
1193 y(training)10 b(examples,)h(that)g(is,)h(examples)e(in)h(whic)o(h)
98 1243 y(the)17 b(a)o(v)n(ailable)d(training)h(output)h(ma)o(y)e(b)q
(e)j(only)e(a)98 1293 y(lo)o(w)o(er)e(b)q(ound)g(on)g(the)h(\\righ)o(t)
e(answ)o(er.")18 b(This)13 b(ap-)98 1343 y(proac)o(h)h(is)f(augmen)o
(ted)g(with)g(a)g(feature)h(selection)98 1392 y(metho)q(d)i(that)g(c)o
(ho)q(oses)i(an)e(appropriate)g(feature)98 1442 y(set)e(within)f(the)h
(con)o(text)f(of)g(the)h(linear)f(program-)98 1492 y(ming)k
(generalizer.)33 b(Computational)16 b(results)j(in)98
1542 y(the)10 b(\014eld)g(of)f(breast)i(cancer)g(prognosis)e(are)h(sho)
o(wn.)98 1592 y(A)18 b(straigh)o(tforw)o(ard)f(translation)f(of)h(the)h
(predic-)98 1642 y(tion)g(metho)q(d)g(to)h(an)f(arti\014cial)g(neural)h
(net)o(w)o(ork)98 1691 y(mo)q(del)12 b(is)i(also)f(prop)q(osed.)15
1860 y Fi(1)56 b(INTR)n(ODUCTION)15 1965 y Fj(Mac)o(hine)16
b(learning)e(metho)q(ds)h(ha)o(v)o(e)g(b)q(een)i(successfully)f(ap-)15
2015 y(plied)f(to)g(the)h(analysis)f(of)f(man)o(y)g(di\013eren)o(t)i
(complex)e(prob-)15 2065 y(lems)j(in)h(recen)o(t)i(y)o(ears,)g
(including)d(man)o(y)f(biomedical)g(ap-)15 2115 y(plications.)i(One)d
(\014eld)g(whic)o(h)f(can)g(b)q(ene\014t)i(from)c(this)i(t)o(yp)q(e)15
2165 y(of)f(approac)o(h)h(is)g(the)h(analysis)e(of)g
Fh(survival)h Fj(or)g Fh(lifetime)g(data)15 2214 y Fj(\(Lee,)20
b(1992;)f(Miller)e(Jr.,)i(1981\),)g(in)e(whic)o(h)i(the)f(ob)r(jectiv)o
(e)15 2264 y(can)12 b(b)q(e)g(broadly)g(de\014ned)g(as)g(predicting)g
(the)h(time)d(of)i(a)f(par-)15 2314 y(ticular)17 b(ev)o(en)o(t.)29
b(In)17 b(this)g(w)o(ork)g(w)o(e)g(are)h(concerned)h(sp)q(eci\014-)15
2364 y(cally)11 b(with)g(prognosis,)h(that)g(is,)f(predicting)i(the)f
(course)h(of)e(a)15 2414 y(disease.)18 b(This)11 b(pap)q(er)g(in)o(tro)
q(duces)h(a)e(new)i(prognosis)e(predic-)15 2463 y(tion)15
b(metho)q(d,)h(based)g(on)g(linear)g(programmi)o(ng)d(\(Dan)o(tzig,)15
2513 y(1963\),)h(whic)o(h)h(can)h(b)q(e)f(implemen)o(ted)e(using)i(a)g
(bac)o(kpropa-)15 2563 y(gation)c(neural)i(net)o(w)o(orks)g
(\(Rumelhart)e(et)i(al.,)e(1986\))g(based)15 2613 y(on)20
b(the)i(same)d(ob)r(jectiv)o(e.)39 b(These)22 b(metho)q(ds)e(are)i
(applied)15 2663 y(to)16 b(breast)i(cancer)g(prognosis,)e(predicting)h
(ho)o(w)f(long)g(after)15 2712 y(surgery)i(w)o(e)f(can)h(exp)q(ect)g
(the)g(disease)g(to)f(recur.)29 b(W)m(e)17 b(also)15
2762 y(describ)q(e)c(a)f(feature)g(selection)g(metho)q(d)f(whic)o(h)g
(has)h(b)q(een)h(in-)15 2812 y(corp)q(orated)18 b(in)o(to)f(our)g
(linear)f(mo)q(del,)g(and)h(sho)o(w)g(that)g(the)1065
866 y(ev)n(aluation)11 b(of)g(the)i(e\013ectiv)o(eness)h(of)e(the)g(v)n
(arious)f(input)h(fea-) y(tures)j(is)f(medically)e(imp)q(ortan)
o(t.) y(In)17 b(applying)f(inductiv)o(e)i(mac)o(hine)e
(learning)g(to)i(prognosis,) y(w)o(e)c(\014rst)g(note)g(that)f
(this)g(problem)f(do)q(es)j(not)e(\014t)g(in)o(to)g(either)1065
1090 y(of)c(the)i(classic)f(paradigms)e(of)h(classi\014cation)g(or)h
(function)f(ap-) y(pro)o(ximation.)15 b(While)c(a)h(patien)o
(t)g(can)g(b)q(e)g(classi\014ed)h(\\recur")
y(if)h(the)h(disease)g(is)f(observ)o(ed,)h(there)h(is)e(no)g(real)h
(cuto\013)f(p)q(oin)o(t) y(at)d(whic)o(h)f(the)i(patien)o(t)f
(can)g(b)q(e)g(considered)h(a)f(non-recurren)o(t)1065
1290 y(case.)19 b(The)12 b(data)h(are)g(therefore)h Fh(c)n(ensor)n(e)n
(d)e Fj(in)g(that)h(w)o(e)f(kno)o(w) y(a)f(time)e(to)i(recur)i
(\(TTR\))d(for)h(only)f(a)h(subset)h(of)f(patien)o()g(for)1065
1389 y(the)k(others,)g(w)o(e)f(kno)o(w)g(only)f(the)i(time)e(of)h
(their)g(last)g(c)o(hec)o(k-) y(up,)21 b(or)f(disease)h(free)g
(surviv)n(al)e(time)f(\(DFS\).)i(In)g(particu-)
y(lar,)c(recurrence)j(or)d(surviv)n(al)f(data)g(is)h
Fh(right)g(c)n(ensor)n(e)n(d)p Fj(,)g(i.e.,)
y(the)e(righ)o(t)g(endp)q(oin)o(t)g(\(recurrence)j(time\))12
b(is)i(sometimes)e(un-) y(kno)o(wn,)g(since)h(some)f(patien)o
(ts)h(will)e(inevitably)g(mo)o(v)o(e)g(a)o(w)o(a)o(y)m(,)1065
1638 y(c)o(hange)j(do)q(ctors,)h(or)e(die)h(of)g(unrelated)g(causes.)
y(Problems)e(in)o(v)o(olving)f(censored)k(data)e(are)g
(common)d(to)j(sev-) y(eral)i(\014elds.)22 b(In)15
b(engineering,)h(one)f(migh)o(t)e(b)q(e)j(in)o(terested)g(in)1065
1813 y(the)i(surviv)n(al)e(c)o(haracteristics)j(of)e(electronic)h(comp)
q(onen)o(ts,) y(while)c(so)q(ciologists)g(migh)o(t)f(consider)
j(what)e(factors)i(lead)e(to) y(long-lasting)k(marriages.)37
b(Ho)o(w)o(ev)o(er,)22 b(the)f(application)e(of)1065
1962 y(mac)o(hine)12 b(learning)g(metho)q(ds)g(to)h(these)h(problems)e
(has)h(b)q(een) y(rare.)27 b(Sc)o(henone)18
b Fh(et)g(al.)26 b Fj(\(Sc)o(henone)19 b(et)e(al.,)f(1993\))g(used)h(a)
y(self-organizing)d(neural)i(net)o(w)o(ork)g(to)g(\014nd)g
(classes)h(of)e(cases) y(with)c(similar)d(exp)q(ected)13
b(recurrence)h(times.)i(Ho)o(w)o(ev)o(er,)11 b(they)1065
2161 y(did)h(not)g(directly)g(address)i(the)f(problem)e(of)g(using)h
(censored) y(data,)18 b(that)g(is,)g(cases)i(whic)o(h)d(ha)o
(v)o(e)h(not)g(b)q(een)h(follo)o(w)o(ed)d(to)
y(recurrence)k(/)c(death.)28 b(Burk)o(e)18 b(\(Burk)o(e,)g(1994\))e
(used)h(arti\014-) y(cial)12 b(neural)h(net)o(w)o(orks)g
(\(ANNs\))h(to)f(approac)o(h)f(prognosis)h(as)
y(a)18 b(separation)g(problem,)f(as)i(w)o(as)f(done)g(in)g(previous)g
(w)o(ork) y(at)i(Wisconsin)f(\(W)m(olb)q(erg)g(et)h(al.,)g
(1992;)h(W)m(olb)q(erg)d(et)j(al.,) y(1994\).)38
b(This)20 b(is)h(done)g(b)o(y)f(c)o(ho)q(osing)g(one)h(or)g(more)f
(end-) y(p)q(oin)o(ts)d(and)g(separating)h(sets)g(suc)o(h)h
(as)e(\\patien)o(ts)g(who)g(re-) y(curred)11
b(in)e(less)h(than)f(t)o(w)o(o)g(y)o(ears.")17 b(The)9
b(w)o(ork)g(of)g(Ra)o(vdin)f(and) y(colleagues)19
b(\(De)f(Lauren)o(tiis)h(and)f(Ra)o(vdin,)g(1994;)h(Ra)o(vdin)1065
2659 y(and)e(Clark,)f(1992;)h(Ra)o(vdin)f(et)i(al.,)e(1994\))g(used)h
(ANNs)h(to) y(generate)c(surviv)n(al)e(curv)o(es,)i(whic)o(h)f
(plot)g(the)g(probabilit)o(y)f(of) y(disease-free)k(surviv)n
(al)e(against)f(time.)19 b(This)14 b(w)o(ork)h(uses)g(the)1065
2809 y(the)i(trained)g(net)o(w)o(ork's)g(output)g(as)f(an)h(appro)o
(ximation)d(of)p eop
%%Page: 2 2
2 1 bop 15 154 a Fj(recurrence)17 b(probabilit)o(y)m(.)f(While)d(their)
h(cum)o(ulativ)o(e)e(results)15 203 y(closely)i(\014t)h(the)g(p)q
(opulation)e(recurrence)k(c)o(haracteristics)g(of)15
253 y(the)11 b(test)h(cases,)g(w)o(e)f(b)q(eliev)o(e)g(that)g(b)q
(etter)i(individual)8 b(predic-)15 303 y(tions)15 b(ma)o(y)f(b)q(e)i
(obtained)f(b)o(y)g(directly)g(predicting)h(surviv)n(al)15
353 y(time.)15 476 y Fi(2)56 b(RECURRENCE)18 b(SURF)-6
b(A)n(CE)99 534 y(APPR)n(O)n(XIMA)h(TION)15 638 y Fk(2.1)48
b(LINEAR)16 b(PR)o(OGRAMMING)124 688 y(F)o(ORMULA)l(TION)15
779 y Fj(W)m(e)i(approac)o(h)g(the)h(prediction)g(of)e(time)g(to)h
(recur)i(\(TTR\))15 829 y(as)h(a)f(function)g(estimation)f(problem,)h
(a)g(mapping)e(of)i(an)15 878 y(n-dimensional)14 b(input)i(of)f
(cytological)g(and)h(other)h(features)15 928 y(to)g(a)f
(one-dimensional)f(time)h(output.)27 b(Complicating)14
b(the)15 978 y(problem)20 b(is)h(the)h(fact)g(that)f(TTR)g(is)h(kno)o
(wn)f(for)g(only)f(a)15 1028 y(subset)g(of)d(patien)o()j(for)e(the)g
(others,)i(w)o(e)e(kno)o(w)g(only)f(the)15 1078 y(time)k(of)h(their)h
(last)f(c)o(hec)o(k-up,)j(or)d(disease-free)i(surviv)n(al)15
1127 y(time)11 b(\(DFS\).)i(The)g(most)e(ob)o(vious)h(approac)o(h)h(to)
g(predicting)15 1177 y(TTR)k(is)f(to)h(use)h(only)f(the)g(uncensored)i
(cases)g({)d(those)i(for)15 1227 y(whic)o(h)f(a)g(TTR)g(is)g(kno)o(wn)f
({)h(and)g(use)h(the)g(input)f(features)15 1277 y(to)j(\014t)g(them)g
(with)f(least)i(squares)g(\(or)f(other)h(least-error\))15
1327 y(pro)q(cedure.)e(Ho)o(w)o(ev)o(er,)11 b(b)o(y)g(exploiting)e(the)
j(straigh)o(tforw)o(ard)15 1376 y(manner)e(in)g(whic)o(h)h
(inequalities)f(are)i(handled)e(in)h(linear)f(pro-)15
1426 y(grammi)o(ng,)f(w)o(e)k(are)g(able)f(to)g(include)g(all)f(a)o(v)n
(ailable)f(cases)k(to)15 1476 y(build)f(a)h(more)f(accurate,)h(robust)h
(predictiv)o(e)f(mo)q(del.)15 1551 y(Our)24 b(solution)f(to)g(this)h
(estimation)d(problem)i(is)g(termed)15 1601 y(the)11
b(recurrence)i(surface)f(appro)o(ximatio)o(n)c(\(RSA\))i(tec)o(hnique)
15 1650 y(\(Mangasarian)g(et)h(al.,)e(1994;)h(Street,)i(1994;)e(W)m
(olb)q(erg)g(et)h(al.,)15 1700 y(1995\).)38 b(RSA)21
b(uses)h(linear)f(programmi)o(ng)d(to)j(determine)15
1750 y(a)16 b(linear)g(com)o(bination)d(of)j(the)h(input)f(features)h
(that)f(accu-)15 1800 y(rately)e(predicts)i(TTR.)d(The)i(in)o(tuitiv)o
(e)e(motiv)n(ation)e(for)j(the)15 1850 y(RSA)g(approac)o(h)f(is)h
(that:)57 1962 y Fg(\017)20 b Fj(Recurrences)e(actually)c(tak)o(e)h
(place)h(at)f(some)f(p)q(oin)o(t)g(in)98 2011 y(time)d(prior)i(to)f
(their)h(detection.)19 b(Ho)o(w)o(ev)o(er,)12 b(the)i(di\013er-)98
2061 y(ence)j(b)q(et)o(w)o(een)g(the)f(time)e(a)h(recurrence)k(is)d
(detectable)98 2111 y(\(actual)d(TTR\))f(and)g(the)i(time)d(it)h(is)h
(actually)f(detected)98 2161 y(\(observ)o(ed)j(TTR\))f(is)f(assumed)h
(to)g(b)q(e)g(small.)57 2252 y Fg(\017)20 b Fj(Observ)o(ed)k
(disease-free)g(surviv)n(al)d(time)h(\(DFS\))g(is)g(a)98
2302 y Fh(lower)d(b)n(ound)h Fj(on)f(the)h(recurrence)i(time)c(of)h
(that)g(pa-)98 2352 y(tien)o(t.)15 2463 y(These)c(assumptions)e(can)i
(b)q(e)f(form)o(ulated)e(in)o(to)h(the)i(follo)o(w-)15
2513 y(ing)e(linear)h(program)e(for)h(a)h(giv)o(en)f(training)g(set:)85
2628 y(minim)o(ize)99 2653 y Ff(w)q(;\015)r(;v)q(;y)q(;z)344
2600 y Fj(1)p 336
v 336 2656 a Fe(m)378 2628
y(e)397 2611 y Ff(T)423 2628 y Fe(y)e Fj(+)502 2600 y(1)p
v 501 2656 a Fe(k)529 2628 y(e)548 2611
y Ff(T)574 2628 y Fe(z)h Fj(+)659 2600 y Fe(\016)p 651
v 651 2656 a(m)693 2628 y(e)712 2611 y Ff(T)738
2628 y Fe(v)67 2764 y Fj(sub)r(ject)k(to)352 2714 y Fg(\000)p
Fe(v)44 b Fg(\024)57 b Fe(M)5 b(w)k Fj(+)h Fe(\015)r(e)g
Fg(\000)g Fe(t)56 b Fg(\024)42 b Fe(y)521 2764 y Fg(\000)p
Fe(N)5 b(w)11 b Fg(\000)e Fe(\015)r(e)h Fj(+)g Fe(r)42
b Fg(\024)h Fe(z)603 2814 y(v)q(;)7 b(y)q(;)g(z)126 b
Fg(\025)43 b Fj(0)937 2760 y(\(1\)) y(The)22
b(purp)q(ose)h(of)e(this)h(linear)f(program)f(is)h(to)h(learn)f(the)
y(w)o(eigh)o(t)16 b(v)o(ector)g Fe(w)h Fj(and)f(the)h(constan)
o(t)f(term)g Fe(\015)r Fj(.)25 b(These)17 b(pa-)1065
253 y(rameters)f(determine)g(a)f(recurrence)k(surface)e
Fe(s)f Fj(=)f Fe(xw)c Fj(+)g Fe(\015)r Fj(,)
y(where)17 b Fe(x)e Fj(is)h(the)g Fe(n)p Fj(-dimensional)d(v)o(ector)k
(of)e(measured)g(fea-) y(tures)20 b(and)f Fe(s)h
Fj(is)f(the)g(surface)h(\(in)f(this)g(case,)i(a)e(plane)g(de-)1065
403 y(\014ned)i(on)f(the)g(feature)h(space\))g(that)f(predicts)i
(recurrence) y(times.)30 b(Here)19 b Fe(M)k Fj(is)18
b(an)f Fe(m)c Fg(\002)f Fe(n)18 b Fj(matrix)e(of)i(the)g
Fe(m)h Fj(recur-) y(ren)o(t)j(p)q(oin)o(ts,)g(with)e
(recurrence)k(times)c(giv)o(en)h(b)o(y)f(the)i Fe(m)p
Fj(-) y(dimensional)16 b(v)o(ector)j Fe(t)p Fj(.)32
b(Similarly)l(,)17 b(the)i Fe(k)g Fj(non-recurren)o(t)1065
602 y(p)q(oin)o(ts)c(are)h(collected)g(in)e(the)i Fe(k)11
b Fg(\002)f Fe(n)15 b Fj(matrix)f Fe(N)5 b Fj(,)15 b(and)g(their)1065
652 y(last)20 b(kno)o(wn)h(disease-free)h(surviv)n(al)e(times)g(are)h
(in)f(the)h Fe(k)q Fj(-) y(dimensional)14 b(v)o(ector)k
Fe(r)q Fj(.)26 b(The)18 b(v)o(ectors)g Fe(y)g Fj(and)f
Fe(z)i Fj(represen)o(t) y(the)e(errors)g(for)f(recurren)o(t)j
(and)d(non-recurren)o(t)i(p)q(oin)o(ts,)e(re-)
y(sp)q(ectiv)o()k(o)o(v)o(erestimating)c(the)i(TTR)g(of)f
(recurrences)k(is) y(considered)e(an)f(error,)h(while)e
(predicting)h(a)f(TTR)h(smaller) y(than)h(an)g(observ)o(ed)h
(DFS)f(is)g(also)g(an)g(error.)35 b(The)19 b(ob)r(jec-)1065
951 y(tiv)o(e)f(a)o(v)o(erages)g(the)h(errors)g(o)o(v)o(er)f(their)g
(resp)q(ectiv)o(e)i(classes.) y(\(Note:)i Fe(e)16
b Fj(is)f(a)h(v)o(ector)g(of)f(1's)g(of)g(appropriate)h(dimension.\))
y(The)e(linear)g(program)e(is)i(implemen)o(ted)e(using)i(the)
g(MINOS) y(n)o(umerical)g(optimization)e(soft)o(w)o(are)k
(\(Murtagh)f(and)g(Saun-) y(ders,)g(1983\).)1065
1225 y(Because)f(recurrences)i(tak)o(e)d(place)f(at)h(some)e(unkno)o
(wn)h(time) y(prior)17 b(to)f(their)i(detection,)g(w)o(e)f(do)
g(not)g(consider)g(underes-) y(timated)f(recurren)o(t)j(p)q
(oin)o(ts)e(to)g(b)q(e)h(as)f(serious)h(of)e(an)h(error)1065
1374 y(as)g(o)o(v)o(erestimated)f(ones.)27 b(T)m(o)16
b(re\015ect)j(this,)e(the)g Fe(v)h Fj(term)e(in)1065
1424 y(the)21 b(ob)r(jectiv)o(e)g(is)f(w)o(eigh)o(ted)g(b)o(y)g(an)g
(appropriately)g(small) y(p)q(ositiv)o(e)10
b(parameter)g Fe(\016)r Fj(,)g(forcing)f(underestimated)i(recurren)o(t)
y(p)q(oin)o(ts)k(closer)h(to)f(the)h(surface.)22
b(Based)17 b(on)d(a)h(p)q(erturbation) y(theorem)i
(\(Mangasarian)f(and)g(Mey)o(er,)j(1979\),)d(for)h(a)f(su\016-)1065
1623 y(cien)o(tly)i(small)e(p)q(ositiv)o(e)i Fe(\016)r
Fj(,)g(that)g(is)g(0)h Fe( 5 Years) s
(Predicted TTR <= 2 Years) s
(2 Years < Predicted TTR )90 2057 y()90 2144 y(:)147
2046 y Fj(\()p Fe(tar)q(g)q(et)e Fg(\000)g Fe(pr)q(edicted)p
Fj(\))513 2031 y Fc(2)513 2057 y(+)680 2046 y Fj(censored)16
b(case)147 2104 y(\()p Fe(pr)q(edicted)10 b Fg(\000)f
Fe(tar)q(g)q(et)p Fj(\))512 2089 y Fc(2)512 2115 y(+)680
2104 y Fj(non-censored)16 b(case)193 2163 y(+)p Fe(\016)r
Fj(\()p Fe(tar)q(g)q(et)11 b Fg(\000)e Fe(pr)q(edicted)p
Fj(\))611 2147 y Fc(2)611 2173 y(+)15 2239 y Fj(where)19
b Fe(e)f Fj(is)g(the)g(error)h(on)e(a)h(particular)f(example,)g
Fe(x)901 2245 y Fc(+)946 2239 y Fj(:=)15 2289 y Fe(max)p
Fg(f)p Fe(x;)7 b Fj(0)p Fg(g)12 b Fj(and)i Fe(\016)i
Fj(is)e(again)e(a)i(small)e(p)q(ositiv)o(e)h(constan)o(t.)15
2364 y(Another)h(imp)q(ortan)o(t)d(researc)o(h)k(direction)e(is)g(the)g
(use)h(of)e(the)15 2414 y(RSA)17 b(metho)q(d)g(to)g(ev)n(aluate)g(the)h
(relativ)o(e)f(e\013ectiv)o(eness)k(of)15 2463 y(di\013eren)o(t)g
(treatmen)o(t)e(strategies.)37 b(By)20 b(building)e(separate)15
2513 y(prognostic)f(prediction)g(surfaces)i(using)d(patien)o(ts)i(who)e
(re-)15 2563 y(ceiv)o(ed)j(di\013eren)o(t)h(treatmen)o(ts,)g(w)o(e)f
(can)f(determine)h(whic)o(h)15 2613 y(regions)11 b(of)g(the)g(feature)h
(space)g(corresp)q(ond)h(to)e(an)g(impro)o(v)o(ed)15
2663 y(prognosis)16 b(for)f(eac)o(h)i(particular)f(treatmen)o(t.)24
b(The)16 b(p)q(ortion)15 2712 y(of)d(the)h(feature)g(space)g(where)h
(one)e(treatmen)o(t)g(is)h(clearly)f(fa-)15 2762 y(v)o(ored)i(w)o(ould)
f(represen)o(t)j(a)e(patien)o(t)g(pro\014le,)f(a)h(set)h(of)e(c)o(har-)
15 2812 y(acteristics)f(whic)o(h)f(w)o(ould)f(indicate)g(use)i(of)e
(that)h(treatmen)o(t.) y(As)j(with)f(all)g(of)g(our)h(medical)e
(applications,)g(the)i(goal)f(is)g(to) y(pro)o(vide)19
b(do)q(ctors)g(and)g(patien)o(ts)g(with)g(predictiv)o(e)g(mo)q(dels)
y(that)f(increase)h(the)g(accuracy)g(with)f(whic)o(h)g
(clinicians)f(can) y(plot)g(the)i(future)f(course)h(of)e
(diseases,)j(resulting)e(in)f(treat-) y(men)o(t)c(decisions)h
(in)f(whic)o(h)g(b)q(oth)g(do)q(ctor)h(and)g(patien)o(t)f(ha)o(v)o(e)
y(the)h(most)f(reliable)h(information.)1065
503 y Fk(Ac)o(kno)o(wledgemen)o(ts) y Fj(The)23
b(authors)g(w)o(ould)f(lik)o(e)g(to)g(thank)h(Don)f(Henson)h(and)1065
636 y(Margarete)h(Mitze)f(for)g(use)h(of)e(the)i(SEER)f(and)g(Guten-)
y(b)q(erg)18 b(databases,)h(resp)q(ectiv)o(ely)m(.)30
b(This)17 b(researc)o(h)i(w)o(as)f(sup-) y(p)q(orted)12
b(b)o(y)e(Air)h(F)m(orce)h(O\016ce)f(of)g(Scien)o(ti\014c)g(Researc)o
(h)h(Gran)o(t) y(F{{003)o(6)f(and)j(National)f
(Science)i(F)m(oundation) y(Gran)o(t)e(CCR{65
953 y Fi(References) y Fj(Aha,)21 b(D.)d(W.,)i(Kibler,)h(D.,)f
(and)f(Alb)q(ert,)j(M.)d(K.)g(\(1991\).) y(Instance-based)h
(learning)e(algorithms.)29 b Fh(Machine)20 b(L)n(e)n(arn-)1065
1144 y(ing)p Fj(,)14 b(6:37{66.) y(Aitk)o(en,)k(R.)f(J.,)h
(Gaze,)h(M.)e(N.,)h(Ro)q(dger,)g(A.,)g(Chett)o(y)m(,)h(U.,)1065
1274 y(and)d(F)m(orrest,)h(A.)f(P)m(.)f(M.)h(\(1989\).)24
b(Arm)15 b(morbidit)o(y)f(within) y(a)19 b(trial)f(of)g
(mastectom)o(y)f(and)i(either)g(no)q(dal)g(sample)e(with)1065
1373 y(selectiv)o(e)k(radiotherap)o(y)g(or)f(axillary)e(clearance.)39
b Fh(British) y(Journal)15 b(of)g(Sur)n(gery)p
Fj(,)e(76:568{571.) y(Burk)o(e,)e(H.)f(B.)h(\(1994\).)g
(Arti\014cial)f(neural)g(net)o(w)o(orks)h(for)f(can-)1065
1553 y(cer)15 b(researc)o(h:)20 b(Outcome)13 b(prediction.)18
b Fh(Seminars)d(in)g(Sur)n(gi-) y(c)n(al)g(Onc)n(olo)n(gy)p
Fj(,)e(10:73{79.) y(Carter,)f(C.)e(L.,)g(Allen,)g(C.,)g(and)h
(Henson,)g(D.)f(E.)g(\(1989\).)i(Re-) y(lation)i(of)g(tumor)g
(size,)i(lymph)e(no)q(de)h(status,)h(and)f(surviv)n(al)1065
1783 y(in)e(24,740)g(breast)h(cancer)i(cases.)j Fh(Canc)n(er)p
Fj(,)13 b(63:181{187.) y(Co)o(x,)h(D.)g(R.)f(\(1972\).)20
b(Regression)15 b(mo)q(dels)f(and)g(life-tables.)1065
1913 y Fh(Journal)f(of)h(the)f(R)n(oyal)h(Statistic)n(al)f(So)n(ciety)p
Fj(,)f(B)h(34:187{202.) y(Dan)o(tzig,)i(G.)g(B.)h(\(1963\).)24
b Fh(Line)n(ar)16 b(Pr)n(o)n(gr)n(amming)g(and)i(Ex-)1065
2043 y(tensions)p Fj(.)g(Princeton)d(Univ)o(ersit)o(y)f(Press,)h
(Princeton)g(NJ.) y(De)g(Lauren)o(tiis,)g(M.)f(and)h(Ra)o
(vdin,)e(P)m(.)h(M.)h(\(1994\).)20 b(A)15 b(tec)o(h-)1065
2173 y(nique)22 b(for)f(using)h(neural)g(net)o(w)o(ork)g(analysis)f(to)
h(p)q(erform) y(surviv)n(al)e(analysis)h(of)g(censored)i
(data.)40 b Fh(Canc)n(er)21 b(L)n(etters)p Fj(,)1065
2273 y(77:127{138.) y(Hart,)13 b(P)m(.)f(\(1967\).)j(The)e
(condensed)h(nearest)h(neigh)o(b)q(or)d(rule.)
y Fh(T)m(r)n(ansactions)j(on)g(Information)g(The)n(ory)p
Fj(,)e(IT-14:515{516.) y(John,)e(G.)e(H.,)h(Koha)o(vi,)g(R.,)g
(and)g(P\015eger,)i(K.)e(\(1994\).)h(Irrele-)
y(v)n(an)o(t)f(features)h(and)f(the)h(subset)h(selection)f(problem.)g
(In)f Fh(Pr)n(o-) y(c)n(e)n(e)n(dings)16 b(of)g(the)g(11th)g
(International)g(Confer)n(enc)n(e)f(on)i(Ma-)
y(chine)e(L)n(e)n(arning)p Fj(,)f(San)g(Mateo,)f(CA.)h(Morgan)f
(Kaufmann.) y(Kaplan,)18 b(E.)f(L.)h(and)f(Meier,)j(P)m(.)d
(\(1958\).)29 b(Nonparametric) y(estimation)16
b(from)f(incomplete)i(observ)n(ations.)28 b Fh(Journal)18
b(of) y(the)d(A)o(meric)n(an)f(Statistic)n(al)g(Asso)n
(ciation)p Fj(,)g(53:457{481.)p eop
%%Page: 9 9
9 8 bop 15 154 a Fj(Kissin,)19 b(M.)e(W.,)h(Querci)h(della)e(Ro)o(v)o
(ere,)i(G.,)e(Easton,)i(D.,)15 203 y(and)e(W)m(estbury)m(,)g(G.)f
(\(1986\).)27 b(Risk)16 b(of)h(lympho)q(edema)d(fol-)15
253 y(lo)o(wing)c(the)j(treatmen)o(t)f(of)f(breast)i(cancer.)k
Fh(British)12 b(Journal)15 303 y(of)j(Sur)n(gery)p Fj(,)e(73:580{584.)
15 396 y(Kittler,)f(J.)g(\(1986\).)i(F)m(eature)e(selection)h(and)f
(extraction.)j(In)15 446 y(F)m(u,)e(Y.)h(.,)f(editor,)h
Fh(Handb)n(o)n(ok)i(of)f(Pattern)g(R)n(e)n(c)n(o)n(gnition)g(and)15
496 y(Image)g(Pr)n(o)n(c)n(essing)p Fj(.)e(Academic)h(Press,)h(New)f(Y)
m(ork.)15 588 y(Lac)o(hen)o(bruc)o(h,)21 b(P)m(.)d(and)h(Mic)o(k)o(ey)m
(,)h(P)m(.)e(\(1968\).)33 b(Estimation)15 638 y(of)12
b(error)h(rates)h(in)e(discriminan)o(t)f(analysis.)k
Fh(T)m(e)n(chnometrics)p Fj(,)15 688 y(10:1{11.)15 781
y(Lang,)e(K.,)g(W)m(aib)q(el,)e(A.,)i(and)h(Hin)o(ton,)e(G.)h
(\(1990\).)k(A)c(time-)15 831 y(dela)o(y)21 b(neural)h(net)o(w)o(ork)g
(arc)o(hitecture)h(for)f(isolated)f(w)o(ord)15 880 y(recognition.)d
Fh(Neur)n(al)c(Networks)p Fj(,)e(3:23{43.)15 973 y(Lee,)23
b(E.)e(T.)g(\(1992\).)38 b Fh(Statistic)n(al)21 b(Metho)n(ds)h(for)f
(Survival)15 1023 y(Data)15 b(A)o(nalysis)p Fj(.)j(John)c(Wiley)f(and)h
(Sons,)f(New)i(Y)m(ork.)15 1116 y(Mangasarian,)d(O.)i(L.)f(and)g(Mey)o
(er,)i(R.)d(R.)h(\(1979\).)k(Nonlin-)15 1166 y(ear)12
b(p)q(erturbation)g(of)f(linear)f(programs.)j Fh(SIAM)g(Journal)f(on)15
1216 y(Contr)n(ol)i(and)i(Optimization)p Fj(,)d(17:745{752.)15
1308 y(Mangasarian,)26 b(O.)e(L.,)i(Street,)i(W.)23 b(N.,)j(and)e(W)m
(olb)q(erg,)15 1358 y(W.)15 b(H.)g(\(1994\).)23 b(Breast)17
b(cancer)h(diagnosis)c(and)i(prognosis)15 1408 y(via)11
b(linear)g(programming)o(.)h(T)m(ec)o(hnical)f(Rep)q(ort)h(94{10,)e
(Uni-)15 1458 y(v)o(ersit)o(y)17 b(of)f(Wisconsin.)25
b Fh(Op)n(er)n(ations)17 b(R)n(ese)n(ar)n(ch)p Fj(,)g(accepted.)15
1508 y(Av)n(ailable)f(from)g(ftp://ftp.cs.wisc.edu/math-prog/tec)o(h-)
15 1558 y(rep)q(orts/.)15 1650 y(Marill,)c(T.)i(and)g(Green,)g(D.)g(M.)
f(\(1963\).)19 b(On)14 b(the)h(e\013ectiv)o(e-)15 1700
y(ness)f(of)f(receptors)i(in)e(recognition)g(systems.)k
Fh(IEEE)e(T)m(r)n(ans-)15 1750 y(actions)g(on)h(Information)f(The)n
(ory)p Fj(,)e(9:11{17.)15 1843 y(Miller)d(Jr.,)h(R.)e(G.)h(\(1981\).)h
Fh(Survival)h(A)o(nalysis)p Fj(.)g(John)f(Wiley)15 1893
y(and)j(Sons,)f(New)i(Y)m(ork.)15 1986 y(Mitze,)f(M.)f(\(1992\).)18
b(P)o(ersonal)c(comm)o(unicatio)o(n.)15 2078 y(Murtagh,)20
b(B.)f(and)h(Saunders,)h(M.)e(\(1983\).)33 b(MINOS)20
b(5.0)15 2128 y(user's)f(guide.)31 b(T)m(ec)o(hnical)18
b(Rep)q(ort)h(SOL)g(83.20,)e(Stanford)15 2178 y(Univ)o(ersit)o(y)m(.)12
b(MINOS)f(5.4)e(Release)i(Notes,)h(Decem)o(b)q(er)e(1992.)15
2271 y(Ra)o(vdin,)15 b(P)m(.)h(M.)h(and)f(Clark,)g(G.)g(M.)g(\(1992\).)
25 b(A)17 b(practical)15 2321 y(application)g(of)g(neural)i(net)o(w)o
(ork)f(analysis)g(for)g(predicting)15 2371 y(outcome)e(of)h(individual)
e(breast)j(cancer)g(patien)o(ts.)27 b Fh(Br)n(e)n(ast)15
2420 y(Canc)n(er)15 b(R)n(ese)n(ar)n(ch)g(and)g(T)m(r)n(e)n(atment)p
Fj(,)e(22:285{293.)15 2513 y(Ra)o(vdin,)j(P)m(.)g(M.,)h(T)m(andon,)g
(A.)g(K.,)g(Allred,)g(D.)f(C.,)h(Clark,)15 2563 y(G.)12
b(M.,)g(F)m(uqua,)g(S.)g(A.)g(W.,)g(Hilsen)o(b)q(ec)o(k,)h(S.)g(H.,)f
(Chamness,)15 2613 y(G.)17 b(C.,)h(and)g(Osb)q(orne,)i(C.)e(K.)g
(\(1994\).)30 b(Cathepsin)18 b(D)g(b)o(y)15 2663 y(w)o(estern)h
(blotting)d(and)h(imm)n(unohisto)q(c)o(hemi)o(stry:)22
b(Failure)15 2712 y(to)9 b(con\014rm)g(correlations)h(with)f(prognosis)
h(in)f(no)q(de-negativ)o(e)15 2762 y(breast)18 b(cancer.)30
b Fh(Journal)17 b(of)h(Clinic)n(al)f(Onc)n(olo)n(gy)p
Fj(,)h(12:467{)15 2812 y(474.) y(Rumelhart,)12
b(D.)i(E.,)g(Hin)o(ton,)f(G.)g(E.,)h(and)g(William)o(s,)e(R.)h(J.)1065
203 y(\(1986\).)47 b(Learning)23 b(in)o(ternal)h(represen)o(tations)h
(b)o(y)f(error) y(propagation.)34 b(In)20 b(Rumelhart,)e(D.)h
(E.)h(and)f(McClelland,) y(J.)24 b(L.,)j(editors,)g
Fh(Par)n(al)r(lel)d(Distribute)n(d)g(Pr)n(o)n(c)n(essing)p
Fj(,)i(v)o(ol-) y(ume)13 b(1,)g(c)o(hapter)i(8.)e(MIT)h(Press,)
h(Cam)o(bridge,)d(MA.) y(Sc)o(henone,)27 b(A.,)f(Andreucci,)h
(L.,)e(Sanguinetti,)h(V.,)f(and) y(Morasso,)16
b(P)m(.)e(\(1993\).)22 b(Neural)16 b(net)o(w)o(orks)g(for)f(prognosis)g
(in) y(breast)c(cancer.)i Fh(Physic)n(a)f(Me)n(dic)n(a)p
Fj(,)f(IX\(Supplemen)o(t)f(1\):175{) y(178.)1065
668 y(Street,)17 b(W.)e(N.)g(\(1994\).)23 b Fh(Canc)n(er)16
b(Diagnosis)i(and)f(Pr)n(o)n(gno-) y(sis)d(via)g(Line)n(ar-Pr)n
(o)n(gr)n(amming-Base)n(d)g(Machine)h(L)n(e)n(arning)p
Fj(.) y(PhD)h(thesis,)h(Univ)o(ersit)o(y)e(of)h
(Wisconsin-Madison.)23 b(Av)n(ail-) y(able)f(as)g(Univ)o(ersit)
o(y)g(of)f(Wisconsin)h(Mathematical)e(Pro-) y(grammi)o(ng)9
b(TR)j(94{14)e(from)h(ftp://ftp.cs.wisc.edu/math-)1065
917 y(prog/tec)o(h-rep)q(orts/.) y(Street,)26
b(W.)21 b(N.,)j(W)m(olb)q(erg,)f(W.)e(H.,)j(and)e(Mangasarian,)1065
1050 y(O.)e(L.)f(\(1993\).)36 b(Nuclear)21 b(feature)g(extraction)f
(for)g(breast) y(tumor)e(diagnosis.)32 b(In)19
b Fh(IS&T/SPIE)h(1993)h(International) y(Symp)n(osium)15
b(on)g(Ele)n(ctr)n(onic)e(Imaging:)19 b(Scienc)n(e)c(and)g(T)m(e)n(ch-)
y(nolo)n(gy)p Fj(,)k(v)o(olume)c(1905,)i(pages)h(861{870,)f
(San)g(Jose,)i(Cali-) y(fornia.) y(W)m(olb)q(erg,)h
(W.)f(H.,)i(Bennett,)i(K.)c(P)m(.,)i(and)e(Mangasarian,)1065
1382 y(O.)c(L.)g(\(1992\).)22 b(Breast)17 b(cancer)g(diagnosis)e(and)g
(prognostic) y(determination)c(from)g(cell)h(analysis.)j(Man)o
(uscript,)e(Depart-) y(men)o(ts)d(of)h(Surgery)g(and)g(Human)f
(Oncology)g(and)h(Computer) y(Sciences,)k(Univ)o(ersit)o(y)f
(of)f(Wisconsin.) y(W)m(olb)q(erg,)20 b(W.)f(H.,)h(Street,)i
(W.)d(N.,)h(Heisey)m(,)h(D.)e(H.,)h(and) y(Mangasarian,)j(O.)g
(L.)f(\(1995\).)43 b(Computer-deriv)o(ed)22 b(n)o(u-)1065
1714 y(clear)e(grade)h(and)e(breast)i(cancer)h(prognosis.)36
b Fh(A)o(nalytic)n(al) y(and)16 b(Quantitative)e(Cytolo)n(gy)h
(and)h(Histolo)n(gy)p Fj(.)h(accepted.) y(W)m(olb)q(erg,)j(W.)
f(H.,)h(Street,)i(W.)d(N.,)h(Heisey)m(,)h(D.)e(H.,)h(and)1065
1897 y(Mangasarian,)25 b(O.)e(L.)h(\(accepted\).)49 b(Computer-deriv)o
(ed) y(n)o(uclear)24 b(features)h(distinguish)f(malignan)o(t)d
(from)h(b)q(enign) y(breast)15 b(cytology)m(.)i
Fh(Human)e(Patholo)n(gy)p Fj(.) y(W)m(olb)q(erg,)23
b(W.)f(H.,)h(Street,)j(W.)21 b(N.,)j(and)e(Mangasarian,)1065
2130 y(O.)12 b(L.)g(\(1993\).)j(Breast)f(cytology)e(diagnosis)g(via)f
(digital)g(im-) y(age)f(analysis.)i Fh(A)o(nalytic)n(al)f(and)
i(Quantitative)f(Cytolo)n(gy)f(and) y(Histolo)n(gy)p
Fj(,)i(15\(6\):396{404.) y(W)m(olb)q(erg,)23
b(W.)f(H.,)h(Street,)j(W.)21 b(N.,)j(and)e(Mangasarian,)1065
2362 y(O.)12 b(L.)g(\(1994\).)i(Mac)o(hine)e(learning)g(tec)o(hniques)h
(to)f(diagnose) y(breast)g(cancer)g(from)e(image-pro)q(cessed)
h(n)o(uclear)g(features)i(of) y(\014ne)h(needle)h(aspirates.)k
Fh(Canc)n(er)c(L)n(etters)p Fj(,)d(77:163{171.)p eop
userdict /end-hook known{end-hook}if