From a576b20f4b93e1ffa315f484e6b3cd83c8f976ac Mon Sep 17 00:00:00 2001 From: Davis King Date: Sun, 2 Jun 2013 19:39:59 -0400 Subject: [PATCH] Added a flow chart diagram to help users pick which machine learning tool to use. --- docs/docs/ml.xml | 30 +- docs/docs/ml_guide.dia | Bin 0 -> 11398 bytes docs/docs/ml_guide.svg | 3713 ++++++++++++++++++++++++++++++++++++++++ docs/makedocs | 2 + 4 files changed, 3726 insertions(+), 19 deletions(-) create mode 100644 docs/docs/ml_guide.dia create mode 100644 docs/docs/ml_guide.svg diff --git a/docs/docs/ml.xml b/docs/docs/ml.xml index c09b9c7bf..5c471b209 100644 --- a/docs/docs/ml.xml +++ b/docs/docs/ml.xml @@ -9,25 +9,9 @@

-

- This page documents all the machine learning algorithms present in - the library. In particular, there are algorithms for performing - classification, regression, clustering, sequence labeling, assignment learning, - rank learning, graph segmentation, object detection, anomaly detection, - and feature ranking, as well as algorithms for doing more - specialized computations. -

- -

- A good tutorial and introduction to the general concepts used by most of the - objects in this part of the library can be found in the svm example program. - After reading this example another good one to consult would be the model selection - example program. Finally, if you came here looking for a binary classification or regression tool then I would - try the krr_trainer first as it is generally the easiest method to use. -

- +

- The major design goal of this portion of the library is to provide a highly modular and + A major design goal of this portion of the library is to provide a highly modular and simple architecture for dealing with kernel algorithms. Towards this end, dlib takes a generic programming approach using C++ templates. In particular, each algorithm is parameterized to allow a user to supply either one of the predefined dlib kernels (e.g. -

Paper Describing dlib Machine Learning

 Davis E. King. Dlib-ml: A Machine Learning Toolkit. 
@@ -470,6 +453,11 @@ Davis E. King. rls object as it
+                  is optimized for this case.
+               

          
 
          
@@ -554,6 +542,10 @@ Davis E. King. kcentroid object.  
+                

+                   If you want to use the linear kernel (i.e. do a normal k-means clustering) then you
+                   should use the find_clusters_using_kmeans routine.
+                

          
 
          
diff --git a/docs/docs/ml_guide.dia b/docs/docs/ml_guide.dia
new file mode 100644
index 0000000000000000000000000000000000000000..9c5de0d073697f0b5c714ff7f0c54db5419ced3a
GIT binary patch
literal 11398
zcmV;1EP2x(iwFP!000021MPila~nx=?RWkPcpP!VjsR;apH*de*RdOZ+zoH+IK1O<
z_>%$>F%mr#DFcwS@`wA|Z)O9eUIf%acYy*`+?bU>ws5+@>XRq4^6~fo@RyfYlQ)ay
z>hk9LH$Om#KTHjq4
z;l+CW>hFI1@$K8U;p*LLzP?$8tIOA6wfOP>&9AQJKiWk<{_xwg$>hT~xR|f!{=MyY
z=j-+I^6BgKVsbryx%kZwPv_5meZIVTeSPu6=Bn*=&u*@6mXkO0tKalmVG`~n-2qnQ1LH?EEF-(
z?dS6Ka8Hj5cXeF2<#FLIS5IEuEZ58V<@$5Tr#ClOi~02ib-jGOI2>p7Y<^{1Y~Qlr
z@OS^^a=pI!!u|g;zgqQKz~dG~jf
z-u=zx>hkH;;=a4Tyk3v?_V31d`%iy69fQCB=L2`c=G}Ihh>N_Cg?znenGVzK(d
zHv9OR9q;sFd)bd)`qV$YuI~r4`PL@}yMwrzzgsM~Z~y!Erh2md1#cN=GJm$dd~^Hi
zgU{sV>3=Vtt+$K)=c}8y&tA-z>&f}#pRTShU)iJnUW}L&xxDzz4?o3^Ke12OXzyrb
z_}NGH>gH}=Aq6UVdQJf`X8!ck_?%02!90EXDaz6oGrxX*wYWpV4H7xh3{yy4(lmyW
zoca%gUR*6+K6!Sty#8E0jJ?VZsw0}uS9(h_98f^B3PeE6l7}qo#p3e$i}gM*h%r+k
zw*N>A9ESDuW_ht#eraVfVBCD!Z!sbv9{4e?UfjHWa*vSxvW5N{lvneMi;u1IXX_y$
z5uy^MkSUII;IC@E_{(R%qfeRZlWTT;GM#4q=(*kbN+tzsMoWP0@Z^~=TmztqL`>zDUPiG5Q7yF<)MoLHpUChYwuqtu=$e)+`4|#8aS6QV+spNf$=PylKfkiiZ*{qP^6SF3{#CP_wJoOx
zHJ7$bZO@ovHdO96_5C&RBpR&1V~C}ssmbCT6I8TvP(s-tfg`yBkLIgQb0GW5Rl`YR
zEV7@4%PAGJ*+eK3fw*cosR*hnLR5+xxmmc!9#8tW<>KP<+1ma-nNQCA)0r42cIQfy
ztNGK#)ju?ISKU$M5Gkjr1Q8qx*R0*wSUVnS?Er%*6>(>W3IU{=_RgHQEf#lJGjLOv
zWXd6>s>dfIEwhislaO*Zo~_J6L{vkAo5OkHhL2NJN5Od*JEmS*aXbgUwS%^U>tJ!pC^A!PxoY2|LA
zG*Gs`&FmY+{2#|wLZuU+n4cPxRA-W%JF*?Gd~?SuUuGyB9CpGWTKs?`p+c^CxN?UC
zriAkh$<#L_d!19({O!}zfBn;+8$VTd>UZtOMo-`L>F@4XK^iQgBoxg&Eg_XsHBewa
z^ae-`$O$R};MBei1+i+JV1y<12@+^1jM?lzi)4(5scMh_3mQ<$NfPM6DV;4yaFSv1
zvuE@5;`z<;(pJ<$;zv^VC9vmfcOlPMTAbiGae_Qpg+oA%Jk2JF4}H%2iSRoh$XI6R
z`$YU@p~x*l&^ToGUTgyVLN8F9#z$tAerpnQ=p
zr>2at$UYj}o-4`;Fp~Ol9j3|$b!6^zL6efvKshxH{#|Ksvu2mquWw$jCjaw#Zhv3D
zYcjmL2_@S&T3m(aNU5ptBNagnujws82+FjW92Txx1$Bo+99d?zV+4H5#1QjIqysc1
zUSEmt+Kx?%pQsdv!G)R}O3aK?Ki$Wwp#qcMw@BHLX5a^RPEishWv!{jM_FXQG;zSp
zL^)ys6&mDXd`XK_Fb%FH+Or%f<{YJI@kg%3|M6z>?&kI6#r)0Utc!s!E+*diq-pT#
z#xn;_s9hN`myAt>A4!A{F8BowNdAk%Af;+6f)9{lM46doW6GW1mZasQG#c5r8=D3{
z!E`yG5i)woD{NU9bFQ1uYe45Q0DIte)@?wPsdop^lv}=B>U
zwjaEU2nUJ`H4p0?i?>+pShI5zP+U+qJxmTp?-wK(X{mj}1Yi#~XQsqu`&BhepdWKc
z3lluHekrRrw!iN>rkr&xQ|{@Q;w9y0FD_T>8*iBU#CE7vPyilDIWUCFN8zecTJ}wpgjdn)%s+vR32udjfrW7-@<|>aqYTnhAWBk{%KhIyjx>~H7m8)*$ycd+B7nn!Z
z()#uuD_TCde$(jQXuqUQs3;p7S9SS5L^Dn{v>66f7M!}R%fcrRnDTG@>8GF1{`04Q
z|JOhK>0cVXRHv7Ho3ur$PBD-^xT^D<1E`>F3C}Ols@Zn$u`n(2D5K|Bm)DE=UEA?y8S4kaT{{s@Ar$W%)j|X#hX{VZ_}|OL@{hKg@3UiN
zgf1#w3qcIPSh%Vc^JC%cR*NNSnss$hd1}&1C>@CWbn(AWy44)C#(O)n<|+BPHSN95
z{R85X?L8?cEqkwhA4Nlfa$t)5#!^4={ju6DpWclAt#0Z9!i6fa)-IAgnds5WzmW!^Ic)fgQug}&uljZ#S*GcyaoSdz0UM|)z><6u~
zxPEsZ@`eagj3EiMA&W;!mBPWLdn6$x4`7=W;#f1^;0`H;!Xh)$P?Dt12o-9n28~Ac
zZO10bPjSxq;L<(g26Nnb3MDA!npvE-;ck({ki73fNz7B9y>KvcKRG_iBHM`bmUPfY
zoPVF8c5hg|5#Jf972_RCe9H*IrU;8^ZGsV|{XbS|CXRN^k#@=3vsA_%i7%S`-2S2oTT1)5sZqs=^
zl-*K!k7fMW{Z$`z0Y(YbH1v@)^x%>?@`ed+c9dgNm}o@
z#v^<8V-wM*nD>^4N3f+7pecIGWu>|i>K#hSXzKDyiz15KP0%s$z(m
zI~A1xY9<61A9vpYsR7NvI||y%5&%hRMnOkeV!s}VtcjT<8RIiUuGu4z1KDDdQk0T`
z<5=pwg2so3#;^*c#!tv%lk?{e7HOchhoBE|ZEnvK~AxwjCKhHW9sRKei~>3Fh6=;F>)2X%(gT
zu?(ryJTf#S*&?AK&3p%{zxEw@C~`k_KFT6{=H8_-6m!eO`)<{EYpZ!=Wz8XGE*_8;
zjV&4PSjw8mh$?v}IIl#CRlC=JfP}Hg402#f-Rp8_u&33tJThIqYd1DseWLxHXmH7$
zE8Q6L^xSiwVy@YS4t+S{21*^uW8Wz1qghLgrD|czILqvJrdvc1F?&EmAc9tneoHPT
zCP}dXL(b6j_V;k6EB`uaXSzpI-{QeZo}Q;*aD|)vKGLDCjv&7?ps_xXk4(vR^>@03
z5`&uAuA!2XeFJ=CO0ImeJHk5qbxh%GT=~9Z3b)`<3m)|k9?f6Ayt#hzYI*bDo5R^=
zy{cO;2qNO#fB`>hZPBb()!@BctcPe1=Jjs}%2a`gIv2_P;Wb%bA+2r_%
z%5ffEpldFg68dpP$_GowemrI3dkd8vYF0u?J|5hN3o_)oLGMx4*+-7&07-!&ATjz$
zD^;D^BM}L(h^R)%hjyK4a{Pp+_H?3r(#`5woy}L1`Ss1q`PI$q)#U8WDon0#miDr#
z_38#zoLZ>3%O5GX%**4>Gv_g?dcj+Motsc9v1&KOVp6P-x?=|=s%dr)7cq6%tcQm0
z`;9c;xy*ZuYcgf&vaSI0-nr%}#Eqva`>wisNbEwQL&sXQiRNvhdCv|yl!uHoH-%lh
z|IYpSEIhbS^W2B%3_D095VoRiz$G6-|*wdBt^j%tZKoks==y#;iewJ
zDykB!`f|Pu53ei^lzo`OrsP&@Hb1z7D)xAbit(8g&<>j^gS32^rjqt<$(Ap3s`)ZH
zTzEy3AufhE)TJp7^Tm3y$y;CjRWq8kjb;i&Al*Tn7m~KP!cpT2>4f465=5C&XU|e3
zoM|W=yQo4+h4VKORj_NlRLqmQs3f2-WWcCVWEU=QQ?vity9b3NmKIrPk%ezLvY-zY
zSr}Y!YR@ee>fAVcIH~4XzB{NA&}c@5{*v87SReoOu7mW=FH|R)nGK
zzxab}A38X3u*t9{!}`iF8YIK;fn=BrE;RHC0utP|mV!fD!Sy4z53zC-_W8((Fm1jQ
z?U=`!>*7jS53$k3mt1_`%$o~iI%{JqxUVF
z7-gxwraZdHRJ^O2Xj2|r*~HMYiD$QyWiFmw+*MBe&1$ijyf3Ke{{Qk9bK6!=7JmuP
zUcEY7{H04XoxECh(;xq~8QJ1Fj{&qqWF(@!1Y>u%|_gtr47&cH$saXNTDNHI4l9S40F$#@@oKSKa%&x%{U0fn0(^^7pZH%##m+_
z`{ik@o;jP)9tKUNW?u#l8KOWYSKj--bqIZsAQT#-By;bH3$+;3v6Oip
zp*LuRoUjW50b||Ppc|+X1U#)1ywG?9jjHX4w=v1y4cWB$w_FFfi=oOG-HSz1nQGu%
zYBSUZN`fJAM^Pk%j;#!brkl0!!Qb!K2p8Hk8cNqkCSiy?d;BfNjTzIi*b
zkMZo~{QBMG*?hftezUw>tlDhAW3*h?XuAfdXgdg?#SM;H1EYiMLbKOnDouUXN31$a
z3k->iSsf#4E->>VO{3~5t9ofwFFmVX0ESe(=tGSM)N#fGdXryMw~{f}o({N$@;+#O5GXN=KjeHc
z^jFLV9K6haA|8u;_*myQ8n7J%9xmFi7jI3cKY4cbdbRdEyye-m7iL~oW?)`iu3i}<
zG}(TP!#*#afLo=W*R$EW6fuf5Ci)So~y&i|oJF{7{md
zeGQpNzSM~2_F>5fo7fhSY7wcP5h)$isb8zl0NhV*a#G}$&HgcG4^0ad@cxS?av2;O
z4lbgpKyDZ!Mkw9s5$~TYa^4jLW%WfNCD?!Moqt;n4vsZ$Yut7mZo`M=HXR<77(+|+
zj7oSvf1@^Mj09=c(E>}}gbJ$8SK7VI9oMo#cU$pG5)Zv!uDTWl{?gEZ#%zt*j>v5E
zZ8O`Bq-7nUpwKi^L4PJwsX8(410*OyGk?Sxklf=-eCRWN%=p*j=bwH$`|}@v`p*{Q
zsvqOpu`OHP@(I>f=or;s07RYqDJenS+KM|U$_QPHPl%gv6JXI+ThS9kNuq8@qC8~|0_U*_Rs&Yum=cDBtlp6-vU1D1
zIZo%8G`RMQVu<$0Pu-2@st3~VkqC;)j38h!^%bHfkQO&+JhE>)HcftlEgjR~N-j(x
zmu_Gg1>jWk2*#k&EfgBmj3j90(hWkxEYg5}#UOk&Pi&)(vLP#7pBO8k#HEDe$X-y;8E)%-o!#*SCs(lqcfKnvF
z8Rh^Pz(?`rBpF-CyK%|B9ogjd37*QO!PPcMK?p^J0uC&eEyk+WZq?mtc2X?snOG(
z?bzt)R1*dUm%Ts^8UcMzMADj_ro2a9hvc=%nrANNaZ|jEA8r2eqm!d7vfsQJMm5u4
zm=%;Mr>Y49QAyd86EupDKnqts@`QmutS0Lhm#fKqvc7z|IA1NiMeLc;?aj59B~4x}
z7Z;aZCCb^W<;_#uMJ;!rzUs%2fxC1~wg<8WIFA(I99$j+K7u%Q<&8>J%cDL(ibgv#
zAAknZy-xWQs-v1bKf+wuUE8tA^Ha?6(DYbeSGU%4d?w8Vk_?eUgZ-WzXXe>6*2-4aqjDaH9GrpGqe=v!#*
zx+kJExKfTGpk|t$Ll8W;30*I{Wl*q%QiqyJG{Y5nN+A`h8d4f(m3<6XaoY?8g?v1$@R^}q69A5vqnMg$bM=mS_>u?L
zYzmnPD(qvTDOF8rG9=j|p&_}=!w7z4mszXYyfK#8uhRe%h?q-K439ye>d6bG&ctd_
zA%%m705qF-DmL%;m)G;!ZDte5Y)DCf=O<^0#h*?jW*Km6p0<-h;KAAkCf$$I|$
z`C{4pzWRP21w}|cpNt;e461J+S>bofXZKk=J)tZGKg~Z!3^4x30gA2
z?tcxp$pjBOAUF@M79Zm=>#Mb+NSc-aF0N<70kcpS0LrOmdep8Z-g6~78RjWOyGhzR
z?blrgM!VE&-Q(VO{lFLXBaoIJ)flZRqwV9Q9_dkBJw0ki!NTCua`ienbE=^PO;v~4
z84_)g#E@o50*iOPfRIw%61`Ct*(-3)p~R1h%Ukh6D_%IHcwx2ppVuDuw!b`gH*el5
z7>*ef&}0uXu%GH>&20+Qu?p}8*PJT`1mrH71G%a_^TEIyBs3(qXGB~)63z!B_q1nY
zEV7q9qYx3BJ!|%CG<()1IoVH}EjyYm%ON2p?mX*Q9Fdip>}1wI~HwEGcB8A@#JRa*YQx12vS!@3?B@DR&pAZ%
z$D24m)@jS&5_0wPd$T4f#G+c$kQov>l7`|)v&5!>Q{Ksl>;W4SFuP?7IysU0Hxp{j%Is2fr{FOa^
zm!0lj+b8;0t&+UD5lk^e$uP|U3uv=}#|=;oA67m}^rUYR9r{U344SD3%xvqJ&x4(T
z%Az%6R6AFB&$XU7UbMRwbJ}~YmS9(sy*V@NyYT%}k`K4VEk@O1R6S!<`=U%eVpRN~
zReMFpV70Elmv)a`%|KNN6gMStoH@^2bgO!s!YoNDQ`)^tyg!&;IB)W<_QulH+kcia
zEg|D4#=J3FW3&g2y3@DKXglUrmoX}kv+1L*2@mDULH7-kDKsYXem@bAx+Xj&p1{27
z|N5stH;3;CjiT@0kB`COdt?Kvd+^iXs$x?>5i)kQQL3q~*4iMUd1#PkY7!(;hs<#3
z$c1C_)1Q87^i-Xmc5TN-Pp3#vWdwRs5%b6}q)fRkJ>4Qf&e`cnId-oL9qJ)9dU|y9
zbpLj2^mKys1aWxb9|WdUJUBy8wYA+H5{HD`RYs_fO*dpt(A>L5PWtK
znU(z54w-Z1!ej^XZj%ymrtG7q9hgdL@~N@Yp3T_E=_JVsM`&pm6T{Rb4VPHzwzS(K
zfp9E7`7BED*Cm3KMo*2N_H4(-PbbMwV@zfg+Yz1~pbVw*WX3I0WG&u5IK@&niN?&$
zyK4lsZ$CDII!S`!5yqus=~^<0XsJ6by#q46q)>d69;PlO$ykb;dTQjfdoMO_I!SIC
zKIhD5G5fep2&xfl4qteW6jgB6O)u!$m1n@zv{R#}ecQ3|(@FA^juCv4oW0*ys(W&G
z3xt6@vB2!40y(a7ltxdDp7!j<#!e^6PJ{Et*)vRHnsUf`@U+%`>gf&%wNRR|cMK$)
zN{FDj$*0Cod-r1_sFNh9;fdohpk$g-0B_Y=GxoGa!XPAtW*%r@a%9ZJEw!NW)4uK4
z_~|71X>jH^Nk~f08jD?KXaTT>P|iJ2rm$mig)a
zk>~(}6UT`{#?%$(iq_1_G$75ArvTnaEYD2H7l~Z~D3@G$BKodX-s@jvyFb58q{;))
z?eVvdKuIv=lrZa|hXkz;!Kw8jKpZj;9kqJ*-`#^e00t+6V+k4|d23wT^i^jF8DuqQ--BBLua|F!8LCu8@rSCS%<77s8
z{Ns&xN;D#Rvp8G7y_vkZT)bUPZhkrI`X^q@*Yh^UqrP@G6$5@Y!@>MIZsV)Q?vl7m
zWe+YeAoOBMop#x@TGQSU_>33u4)>)*qM2=)LeZf!0#_H>yB4}0&fvNMl%Ofg^juBrN~x-|B?u)pS&cy@
z%FKu;XLM*4K2+lL=AGRuy@z;A`ExAYtpXof6Z#dzqY*h}zbyu$LjfD{oEq_v3?ZI`
z55+rtO1uL;Jp;OjS(J5V*d{AzLp%!k8zLTb4I;U_S4Pg&iN}KhNfnqi=iC%eBc49Q
zL;6s}BV(Lq4EHH|3m!4Q${u`kCNg%87He|sWIuo-9BJIY)LMMpn|NYiqZ}kEoO6PY
zLlupA9y9ao^_F^2k37&=Lx90~{G=Y(wFk18ahIAQ9u2q!5)?nsNaA_?DG*jY&KP5b
zeGW4)U@;1P|3z({VH^C^`>f*Gvw1sd&zafFV+0E+rdaz~t3VB=-M;0N;q@ZwU@m-V0-GHgq+040G_L-5;NA#
zN!&o0BhTWY!x&6`blB`QudP(0G1R^t**NM1Yc(*Q(3D++LS*bx4S|@w&A+hvJlleV
zP}_O?no+yzwM;^Oub2=j=W~F0h^qDtAy3TBWr
z*DPTL3g9KHki1(z&kVVs6i0@9Xz#ZExg5LJ`T3=Gqv#`U?M8DtbZA`vIx(N2QM?&6
zCuC-9l=GdWuIiu+viwk#GdyA5_PiD!fE{D5*>%W(eEz6U(#k1&ut@AF>#9=op_8}!
zhp2a~vQC-qS0s)lO3^%0Bb)?Mq?M(9FA`@^Rq8_}tRkIE6}%6f_3H8*SYcC~!_v*h
z<|OdVRKbs;_e_U4?O>%50^f{=>es4`>q-ae371(XVEW)N&L=etSr8NxXt55+lyGvLkc
zGnMSI6lSTJTbKi>8FEK5!;UjU7K!3>fgq;ZI%nrPE79;*WEKKw%F+HW6--$U?O@hO
z=j4RWu)M;D;+(+=&7(LH{dA7`PzlYK@+0tMoP!~mjh~{1a(IlSH|OkH=e?IwMDiQU
znzXjgiGC!7Vj9pVMl#2CI`zH8b9hj#UwkOi$s>@?@F%QyJSUSq_7t=fnIYQ~K+=&;
zR;*7t_pkGw*bj%8+1Qaiai@My6^gMGNC=)DbntkP#yKa~+F(#KCwj0XoN`*6Q^0|e
zD9wy|%%L#K$su#~s;#e~91JB2`i;TC4Nu!4RU?=&3Fe_ZDT*UBZ4#tz@*AYcRSO^}
zY(~ssF3`1XQu83D5^|Dbw2IrY&ihpj#qnFGe$*o>==r`;YVVPaK5Vp
zs`5~6FKBRTF?i2KN?s{sK2FWt;yWY_nPQ$9qSM!bpgUobB2{b3$uk6#AD}}tMWi`IDD23|iMDb@`L1=|Q&4Wi8G-#qH4#@o%K}o6
zVqy^r+O4&M^6x|ve$FvznxANV~oQ53eFJ$`yZBWP%*aB&(_PL*9q5yG0x!RWH$RR3B3fI6l%_1Gb98r
zZpr~M%FO&7v+e5sPBy_)nR0fnvy#8rG6fsKe5`2(l|ST%s`I3BS`6b;WzCZ!HG=UP
zbLqP0WWaBTU&e4Pzowf7ClHECs3j7&xsHu-`Y_HzB?|K3+~iT5hJJ1o294m2hsCQN
zc*f%8p6*v=!c@IQ*RF-$!%rSv@~Bc*cy)w|D~=Ss+P=hMG|&=vzL&(ELCMMKp#sJ{
zI6L{A0%S$g8vzAM&62Bg0B>C3v#O|Ync1IF(1ni+3r>|;XZK3)xfXnW70L7*jUpK;
z_q4!3Xk=3gXh|u?(oV;}mo_?s4yxdxT6gjA9OW_kRQ04C45-;FuY6+l3=A|yLIE71c=#qF5|&7!y4zD8t2F$&XI?zi(on>&N2M}irz2K
z{3&tLk?!m$21Vv?iF9H}COrJk;^BAJ-@}SbTyn~g6tmrBW1Yr2Jy_?#l+NJXV-4zQ
zuRi6eWUNa&6tKiZhGm$Eb>?~S$$PPTDj?L(%GtftT8`(}l${f?eFTw{eMB*{Hs<+G
zI(Q9ou^#FKiw7qlgBTll8?9jeTB#XN)*z{%(+(ySO*4b$n36%Gr0TS@bD_1!xWALd
z8YM@&yO_|^2U=pzcS1UY4zV6;#taYdj+_FhijS$xrE*Jl@tPv_$*~boiS~7=0KTbnEDUT!;Qn7Gf%>g>n37^7NSH-qOWJAYSbdzVhpO`Ad_vv5k~c-CPVd;m
zl{W3Yw0K9s!}EP>wD%fC1AbMVCqOC^vl0?(ZUZ4&gRed<1&6cF=JnP5-D3IMv&{?p
U@A-26^0#OIA9GpE&i5k&0M`P7cK`qY

literal 0
HcmV?d00001

diff --git a/docs/docs/ml_guide.svg b/docs/docs/ml_guide.svg
new file mode 100644
index 000000000..0d2f127d4
--- /dev/null
+++ b/docs/docs/ml_guide.svg
@@ -0,0 +1,3713 @@
+
+
+  
+    
+      
+        image/svg+xml
+        
+      
+    
+  
+  
+    
+    
+      
+      
+    
+    
+      
+      
+      
+      
+      
+      
+    
+  
+  
+  
+  
+  Structured Prediction
+  Markov Random Fields
+  
+  Dimensionality Reduction
+  
+  Regression
+  
+  
+  
+    
+  
+  
+    
+      
+      
+      
+        krr_trainer with
+        radial_basis_kernel
+      
+    
+  
+  
+    
+    
+    
+      Predicting a
+      true or false label?
+    
+  
+  
+    
+    
+    
+      Number of
+      features
+      < 100
+    
+  
+  
+    NO
+  
+  
+    YES
+  
+  
+    
+    
+    
+      Predicting a
+      categorial label?
+    
+  
+  
+    NO
+  
+  
+    
+    
+    
+      Predicting a
+      continuous quantity?
+    
+  
+  
+    NO
+  
+  
+    
+    
+    
+      Do you have
+      labeled data?
+    
+  
+  
+    YES
+  
+  
+    YES
+  
+  
+    
+      
+      
+      
+        svm_c_trainer
+        with radial_basis_kernel or
+        histogram_intersection_kernel
+      
+    
+  
+  
+    
+    
+    
+      < 20K
+      Samples
+    
+  
+  
+    NOT
+    WORKING
+  
+  
+    YES
+  
+  
+    
+      
+      
+      
+        svm_c_linear_trainer
+      
+    
+  
+  
+    
+    
+    
+  
+  
+    NO
+  
+  
+    NO
+  
+  
+    YES
+  
+  
+    
+      
+      
+      
+        krr_trainer with
+        radial_basis_kernel
+      
+    
+  
+  
+    NO
+  
+  
+    
+      
+      
+      
+        svm_rank_trainer
+      
+    
+  
+  
+    YES
+  
+  
+    NOT
+    WORKING
+  
+  
+    
+  
+  
+    
+    
+    
+      < 20K
+      Samples
+    
+  
+  
+    
+      
+      
+      
+        svr_trainer with
+        radial_basis_kernel or
+        histogram_intersection_kernel
+      
+    
+  
+  
+    YES
+  
+  
+    
+      
+      
+      
+        svr_linear_trainer
+      
+    
+  
+  
+    NO
+  
+  
+    
+    
+    
+      < 5K
+      Samples
+    
+  
+  
+    
+      
+      
+      
+        sammon_projection
+      
+    
+  
+  
+    YES
+  
+  
+    
+    
+    
+      Do you have
+      labeled data?
+    
+  
+  
+    YES
+  
+  
+    
+    
+    
+      Are you trying
+      to label things
+      as anomalous
+      vs. normal?
+    
+  
+  
+    
+    
+    
+  
+  
+    
+      
+      
+      
+        svm_one_class_trainer
+        with radial_basis_kernel
+      
+    
+  
+  
+    
+    
+    
+      < 20K
+      Samples
+    
+  
+  
+    
+    
+    
+  
+  
+    YES
+  
+  
+    
+    
+    
+      Go get
+      labels!
+    
+  
+  
+    
+    
+    
+  
+  
+    NO
+  
+  
+    
+    
+    
+  
+  
+    YES
+  
+  
+    
+      
+      
+      
+        svm_c_linear_dcd_trainer
+        (see one_class_classifiers_ex.cpp
+        example program)
+      
+    
+  
+  
+    
+    
+    
+  
+  
+    NO
+  
+  
+    
+      
+      
+      
+        svm_multiclass_linear_trainer
+      
+    
+  
+  
+    
+    
+    
+      Number of
+      features
+      < 100
+    
+  
+  
+    YES
+  
+  
+    
+      
+      
+      
+        one_vs_one_trainer
+        with krr_trainer using
+        radial_basis_kernel
+      
+    
+  
+  
+    NO
+  
+  
+    YES
+  
+  
+    
+    
+    
+      Do you know how
+      many categories?
+    
+  
+  
+    
+    
+    
+  
+  
+    
+      
+      
+      
+        newman_cluster or
+        chinese_whispers
+      
+    
+  
+  
+    
+      
+      
+      
+        kkmeans or
+        find_clusters_using_kmeans
+      
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    TOO
+    SLOW
+  
+  
+    YES
+  
+  
+    NO
+  
+  
+    YES
+  
+  
+    
+      
+      
+      
+        discriminant_pca
+      
+    
+  
+  
+    NO
+  
+  
+    NO
+  
+  
+    NO
+  
+  
+    
+    
+    
+      Is this a time-series
+      or online prediction
+      problem?
+    
+  
+  
+    NO
+  
+  
+    
+      
+      
+      
+        krls or rls
+      
+    
+  
+  
+    
+    
+    
+      Do you want to detect
+      objects in images?
+    
+  
+  
+    
+      
+      
+      
+        structural_object_detection_trainer
+      
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+      Predicting the labels
+      of nodes in a graph?
+    
+  
+  
+    
+    
+    
+      Binary labels on
+      nodes in a graph?
+    
+  
+  
+    
+    
+    
+      A chain structured graph?
+      (e.g. words in a sentence)
+    
+  
+  
+    
+    
+    
+      Are you trying to make
+      a BIO or BILOU tagger?
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+      
+      
+      
+        structural_sequence_segmentation_trainer
+      
+    
+  
+  
+    
+      
+      
+      
+        structural_sequence_labeling_trainer
+      
+    
+  
+  
+    
+      
+      
+      
+        structural_graph_labeling_trainer
+      
+    
+  
+  
+    
+    
+    
+      Trying to solve an
+      assignment problem?
+    
+  
+  
+    
+    
+    
+  
+  
+    
+      
+      
+      
+        structural_assignment_trainer
+      
+    
+  
+  
+    
+      
+      
+      
+        structural_svm_problem
+        (Used to build your own
+        structured precition tool!)
+      
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    YES
+  
+  
+    NO
+  
+  
+    NO
+  
+  
+    YES
+  
+  
+    NO
+  
+  
+    NO
+  
+  
+    YES
+  
+  
+    NO
+  
+  
+    YES
+  
+  
+    YES
+  
+  
+    YES
+  
+  
+    NO
+  
+  
+    NO
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+      Do you have
+      two views of
+      your data?
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+      
+      
+      
+        cca
+      
+    
+  
+  
+    
+    
+    
+  
+  
+    YES
+  
+  
+    NO
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    YES
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+  
+  
+    
+    
+    
+      
+      
+      
+      
+      
+    
+    
+  
+  Classification
+  Clustering
+  
+  
+  Machine Learning Guide
+  
+    
+    
+    
+      Do you just want
+      to visualize your
+      data?
+    
+  
+  
+    
+    
+    
+      Are you trying
+      to rank order
+      something?
+    
+  
+
diff --git a/docs/makedocs b/docs/makedocs
index 7e5d93d86..2a8caf18e 100755
--- a/docs/makedocs
+++ b/docs/makedocs
@@ -152,6 +152,8 @@ makedocs ()
 
     cp docs/*.gif docs/web
     cp docs/*.gif docs/chm/docs
+    cp docs/ml_guide.svg docs/web
+    cp docs/ml_guide.svg docs/chm/docs
     cp -r docs/guipics docs/web
     cp -r docs/guipics docs/chm/docs
     cp docs/*.html docs/web