ai4r 1.0

data/site/build/site/en/machineLearning.pdf

@@ -0,0 +1,337 @@
+%PDF-1.3
+%����
+4 0 obj
+<< /Type /Info
+/Producer (FOP 0.20.5) >>
+endobj
+5 0 obj
+<< /Length 652 /Filter [ /ASCII85Decode /FlateDecode ]
+ >>
+stream
+GauI44d4VC&;GQ1$;.o9#O&c^kMhoQ`'k;]8IB9BR>D=^6C7K)7j$BA/pF(dE9+mKYUo.9?1"Wtma`Oj&LOI/#GhCS"U%g+7ATfXPkT5j6u\N(W&UM)M[ePaU_6>mL8A%^\G.2V;,UCJo2F?FW5jZ)`.Q^a7V]>!hb`kBfeG=:@gXGK"8emGnY^PWG5LaZ.AH8('45aD@2^K<(Z3f]L#[VrU%^7tBk<"(c.-0[.%0q,K_5g:85VGEUV8U4Aj'u)XWNfc\nlS6#lAYHeM)0"7`"C_Rh"@rM$ArXpiNE38o`KJ$!OQY$7*I3`PpsgGh*gO^'j8T`$+'9&f,S!%C`&m_i4B0O!jtX[EZY-"<=5gA`fFKUQjW`Qp^Iq],b&N%tLCU92QLG3"*g:?3tD$JK'p1E_s1+jUh:5F5eJK1iWrN1oQ`Q,WX.T:shB,ch\^BXgB^o9T^\6Hi5E?n\jaZO+m:7MJ&^t7XrB'VDrRX`q*Ftl]!f/%IgK;m;!dSj\[Q.-jVmM`6HX;WKLbHjRVm!'u>s)W=3-G8lD2!?$s=.R)R,PZk#>)X1fIV.m/D>Q:A.2MeLtZ7X9f"#Ni-PSW0]\(7N-B(/n%gO;-2W@4J6BYGEp>H!TRHY!EJV]gX.a@@flD+fsS_FH21~>
+endstream
+endobj
+6 0 obj
+<< /Type /Page
+/Parent 1 0 R
+/MediaBox [ 0 0 612 792 ]
+/Resources 3 0 R
+/Contents 5 0 R
+/Annots 7 0 R
+>>
+endobj
+7 0 obj
+[
+8 0 R
+10 0 R
+12 0 R
+14 0 R
+16 0 R
+]
+endobj
+8 0 obj
+<< /Type /Annot
+/Subtype /Link
+/Rect [ 102.0 532.666 252.992 520.666 ]
+/C [ 0 0 0 ]
+/Border [ 0 0 0 ]
+/A 9 0 R
+/H /I
+>>
+endobj
+10 0 obj
+<< /Type /Annot
+/Subtype /Link
+/Rect [ 102.0 514.466 442.28 502.466 ]
+/C [ 0 0 0 ]
+/Border [ 0 0 0 ]
+/A 11 0 R
+/H /I
+>>
+endobj
+12 0 obj
+<< /Type /Annot
+/Subtype /Link
+/Rect [ 102.0 496.266 201.32 484.266 ]
+/C [ 0 0 0 ]
+/Border [ 0 0 0 ]
+/A 13 0 R
+/H /I
+>>
+endobj
+14 0 obj
+<< /Type /Annot
+/Subtype /Link
+/Rect [ 102.0 478.066 254.312 466.066 ]
+/C [ 0 0 0 ]
+/Border [ 0 0 0 ]
+/A 15 0 R
+/H /I
+>>
+endobj
+16 0 obj
+<< /Type /Annot
+/Subtype /Link
+/Rect [ 102.0 459.866 276.296 447.866 ]
+/C [ 0 0 0 ]
+/Border [ 0 0 0 ]
+/A 17 0 R
+/H /I
+>>
+endobj
+18 0 obj
+<< /Length 1823 /Filter [ /ASCII85Decode /FlateDecode ]
+ >>
+stream
+GauHL>Ar7S'Roe[&GkAj86D%d(YaK4M@OL.[baB6"9iYi10[n'.4JU&rJi-r3csH86dXM:9iC&_Icj&2Qq^o5-9ilBpb$raZlcuP`0=(Gn3-R>r-Eb,I)9.71")Ic5##A--I:P%%]_P4Q4jU\R7#rVSG&oE_Y1d5XduMjN1;mE]HQn=?!\G?</XZ=HrqsXg@<<+-6c*j;<;P@k1hr&5?i\A_lXU#ogfc"]pLW#[ZnMDGUoUB+_S_$>6/@e<=D"[=<Jd;HBZIYbY)@PYEqeBPru3@[rhT@;l0%YjO\%c6>tDV7+B10/7^cdF^sJV'1JU\8gZVmfuenHpeFufO/NdBcB2^pm^uG#;N10"p#[IG>ATK#Sh%'NM-R!rG-mAE0qKmme.VNI]*"p0q,3^TT&EVAYbCVbec",JZr=XBhZiUQ92SYXP0noU9;3IM8=!6@"H08M'M1%2<G=Tb[`kO$\[?i;b8$n/GF'h2.'9MWaS0CW5G@UJAf3qbhg))J+9aThetA4A<;h?1@&JH&(TUg>X<cqHOj#=i7#'j0ghQE_K?Q5BK3'nKJX4FuW0N1OWA$(C3^M*#&6-`$#_(*SU!*jcU];OYHm+Dfa*V6e)8>D$J<1KmB^:DZRCj8;Aa-K47Y'6i5pn@dN>,Npei6>=(2,Y$-nX9gjV0SD3bjj%Jf;3"HW?:Gp[Kqh1tuRQRj6Lm!Th.*\C22P].nW'@7#fW6U2<Tkms(,+"E>,ZMm51aju*U,bVeMKVK[c+Nt@!>/=>KPN%g2PVjphjGI4q`Mg!h=;BcE]kfZ\CflYnmW4$U"#\DWSE@Dqe(tN86DIJ:Sej*oG@gT`/>n;N;/J5F#.:\+LE()K!('g=XW3-?Yh<Q%Z'Vt;b%hB8.Y6MW98n(o/K<_DB4M2YQBu*Con&!QS[.QKm_r%Zd$a24J]#Y13m[5udmbh9"/]:Z>=8k2J"a2B5#=tc_Yr8I&;Llni<0*doC&'r)-d+<f%H`%14)-VLU2NLgd4g'0P[A<d2GcOlb_4Yn@Z?&?%.-Mq;V)E0-Fai)]=,IL0)93\kqMEoGfH=<.6VI:SkH+M*I-B>3c([X``50Sh](c]\2oQ:.U1f:9fmf?I'*QBV5$OqE4G,lg;u)9g?efE8iCtZpq]'K^!D-D&1o%i",PQYLB!4;)W;#)(b(+]NdaeUq/]/4dpWVOb!I%ibqR<*dT(?I_C:6O2U\BZgj_RoH#p@PtNcXH@9Ch4ILg,eL<OZO`=]S^A3#8QOB);LuV:jE*+(I6fn9=["5e0cm.9>oHL"g&%^mW"6u*>jB*L8FfI&6C3F\[$6GrWO5TaL#Tgr-$j\p8"-52VoJZ+L#lG8Tc"-^Y]t-/`^qo?LhebC+4)l&h6nC<@03CIENSAOPlps3??]N2(l3**KHZ>cRM40RqH9P#g\=\<njKX+8kF=qVkW33-hFVPdX.eGH:aPOC1%*7_pVbCMK;D*k\(?$/7m-lf4Kd"jH&Vt^M8N`^)Lp!1>EtliXf+itGBjRYiU1XAgQloPQjVQT-fZh5CERN'o-T2QYDQ5J@gM$P"S_cQk32d;[uR)bdi:Dq29/G`iI:R,0O\8j(@P\K@M7DN-$SWUFN@=^c\Mi%.V(/?4DlJE+;QJm2\LmSm6^6VNFd?adP#e<s*eAmAc3T1\mOg-?N#be2sanel;sLXIP`=D^JbZei7H+'46>,/E]to57]lj\?52gLq[;+bM](tlbki6O,iYN\UGo5)NoA^T5]=\&CFQsHRldCn<HHn#UY:^Zgh4M5]G(j^0@HXg&+__P)VMhnC)8L,bdWdj0R@DC+jI,_ckZLdj]Msn~>
+endstream
+endobj
+19 0 obj
+<< /Type /Page
+/Parent 1 0 R
+/MediaBox [ 0 0 612 792 ]
+/Resources 3 0 R
+/Contents 18 0 R
+>>
+endobj
+20 0 obj
+<< /Length 1771 /Filter [ /ASCII85Decode /FlateDecode ]
+ >>
+stream
+Gau0DD,8nW&H9tYnBhaQ&o>67IRBqE:oHA$lE3N&<<P``]Lf1OG&G!E6eD5(2g0$+)!i$@]guPWGWkNebcX`24(k=Z5Bj%^LVNRFO\8s"pO2fi]]/p@nrUAudE9#Uq0hG%>44L)@4qK9.a"fek=uijdsL7<;I?%d%P8`]Ag!p(#%T\-GcZ/A$V<USdE9!@-8?kpS)eqANALF]/u54D-pePmgP)j)S-.Y#A.F[W_EDI7,+7*hEip`mVK6m%1jHpoJULuC.^dbul7^:*l$Qs9Sh"k=pcNe;5j=[IN^_@H!4C&JFBrt^^iRPC8>/^e92$#j!6\+Ra]-Ql-YBb#DkkK"6MX[No8TRY@DU8Fp7o4.][*#Fnf5!lL^;^Ip;_'W]dP(rpO0A2pE4Qr?BEW"I:5PAn)s=[p_2r5$R-5S7[LD)ZqN/6(bLtAoY&I#O_2TG.a9mCKr'uTQKaYk8VRY42]Abbk(C7Pc,mO0kkPgIWY11F'Gu2/f%F5@X?YQ8gY0Id*4Z3_&kdjfWMNdte-rR*n`)m#OPrJbR:d[1h9"c'Ghl5nq^ABH#%7;Fi+BXRM):fC$'2f[1cEH[cm0Zf+.uhd<roHo4VIQ"2L<`lD/9WN;C>j^48>sK>^3M1JeQMT8eJrSAP=!4MQVP5;9MYQASS>]/c6EI:I8[\i>aK&*7(=[_]08/,-'A18?G'h"\WMTQcPD%L7UP=)b\s69=/Z%/*qek`%'69cVmf^'hA3K=6l^*pTYpKkS#>'dE1dqW#S8#1e[Ec0&lc8+%-*?=6"7Cf?t,rp#`(6Y"!,_m=ds*3+SpgIop/BS7\f:P*;rSeC#HUT^F`g0<2PEb`\eHXhoJ]1`&>#_U_6l'#/(J\(af"%fj?ofcgQ_>Tp10ZFAFESUE)W#N/qqJ63d4ISpfA6,;qoNJH\NEr\2aGe!Y9muiPALu8W39d]$>QY2N(Gjh9V[;IbEmrnFe;[JHIfA/=2KCN7c`#=sRfR;<'rRXf/&C@*omreBL)hd\5IAatF!cSDp&Ue?56_D_+=I8Rok!s&p)qX[M1,IbW@'djghZ5o&jh+O@,R5;2Br):M0A^]3Wr$c\IfGPMX[N\hIN&p17p?i&CeWo?bE:,ZQ4)3Ma7a\JAMIr5A;tm[lYY$YJo%Hlfn5fb1FO-l=<2D@Ms!>Q06UfI]LNel+fpus/nLVY&glob!"n35F'$L8>&IqJ>X<,QE\&ELIhX"bV*rs4E7j@HmIshI@%e\eAH1QFAq;'gA'^_$]d[H1-Lkjo3n^0_Pbbhj9X9UnL23tM'?idE]<eYKBqVeOes(X!Fh:ObN4"E''[I=8^Yk4ihoXB``jOR8QN!;B)Z3-p7gkr`-5-CKF]]^:i,3ORgI8qTW@FcF];F2)l_6>IaQI7QX'^=[1)UTFEU:Z>MZMQKlp"q<]Aq</jQ^Q79!i,7*hSs.?M.;J/.CfEBMKJ:C9,6fLCaeV7s2?Y-aiC:]+/abdR$R])I4M51#CPmGoq@h%k-LmFW^B>-)J_h>EVbgYHq)6"?Kq<lO*`I_:%8bH=P<XHtlISI9(a2TRsld5*uurnH:uVapDY0q;Hsr/s<U.]-&MV4`c@E9'U`gg+_ipC;[!I6UTJef4#N'R(#%HHblVca)Fu%\#i!QJ-93/.HPg^)J)r.$LRI:&gRl8a*t'sVU<=Z=IRQ!0^&VtiK$0A^k]#l9ENA:Nru<5qEZM#70u4F=RLU*7kFn,%kMk"oK:5HI@e5_3R'/9G^HsQ&uB".:odjc3u+M.pb.=#E+f~>
+endstream
+endobj
+21 0 obj
+<< /Type /Page
+/Parent 1 0 R
+/MediaBox [ 0 0 612 792 ]
+/Resources 3 0 R
+/Contents 20 0 R
+>>
+endobj
+22 0 obj
+<< /Length 798 /Filter [ /ASCII85Decode /FlateDecode ]
+ >>
+stream
+Gat%"95g:b&AI`dk3*ej@:(Q*PIccX2,]ndjj<^Om1!CPL(l_/o'\Y&QN[UU[U_,1IZIVKcecUskLf[tq*]9$5h]N"HBD!N#c,7C+j91E.LG2<$4TYsLNaDY<so2=>S',F5(Y6Ypn^eS"io^*`&m`SPN2ct=[b-Yb6r0^A=:1PD"qSeHhXhR%G#i:p*S+%k"oa)^OY2QJ$$neQ&!pO(F0E@ei]^\orQbV=)R!r()dJM#ue&1as.NfXA*NI-_XN/$4h"4&h%N'Y=ONb#5+ac:j)Spk)Xf=CQnDHp='=nBL&dn"![kk4VB/:K^[o)V\Wpi-H3g&[e`.f%WNpCVhE'/l,#f1QHp=YD3f+WEJG#M9&Sb)LdPhN%N+@S&>pPW0^E]7]s1(U]#:FA0I8Q1#;07NKLN^KUE.Vu*Ap`%`hc+pJ5,<f#i,)7i7L%@4T3JUnGG`oE0k$S;J]#L#6Y)AlL_Mgi45C8?nS:m8Z=F13$8M)o=T7^1YQP.>,q(QflYgj5tFScc`l('Xj(TE*;A^'Q82oIAkX=P!eD"13o0TIB5Lsne%@Aic)rdJ[9sOY0cZE_,(q[I'PA_MC@#63g6er[OUD-1<5;o89d`AAZIDI[U4AG@L1_=KHR#ji`c\W^1<*U=s0@#npGf:q,P;dKZ5KUN-aK8?:[LQGD1QPD&rY`mLC!(IC23X;Ui'V8[mAM^-7T=i'!euB1l;P[8oUOSUG^Mur;!HKl)$L'W3AApr*kiCM#"cjd=X&LPM:=jFJ6%S9jDVG2Q2FY2[d=N`BV$_G$k4N<d!pE~>
+endstream
+endobj
+23 0 obj
+<< /Type /Page
+/Parent 1 0 R
+/MediaBox [ 0 0 612 792 ]
+/Resources 3 0 R
+/Contents 22 0 R
+/Annots 24 0 R
+>>
+endobj
+24 0 obj
+[
+25 0 R
+26 0 R
+]
+endobj
+25 0 obj
+<< /Type /Annot
+/Subtype /Link
+/Rect [ 90.0 629.666 259.296 617.666 ]
+/C [ 0 0 0 ]
+/Border [ 0 0 0 ]
+/A << /URI (http://en.wikipedia.org/wiki/Decision_tree)
+/S /URI >>
+/H /I
+>>
+endobj
+26 0 obj
+<< /Type /Annot
+/Subtype /Link
+/Rect [ 262.296 629.666 434.94 617.666 ]
+/C [ 0 0 0 ]
+/Border [ 0 0 0 ]
+/A << /URI (http://en.wikipedia.org/wiki/ID3_algorithm)
+/S /URI >>
+/H /I
+>>
+endobj
+28 0 obj
+<<
+ /Title (\376\377\0\61\0\40\0\111\0\156\0\164\0\162\0\157\0\144\0\165\0\143\0\164\0\151\0\157\0\156\0\40\0\164\0\157\0\40\0\111\0\104\0\63\0\40\0\141\0\154\0\147\0\157\0\162\0\151\0\164\0\150\0\155)
+ /Parent 27 0 R
+ /Next 29 0 R
+ /A 9 0 R
+>> endobj
+29 0 obj
+<<
+ /Title (\376\377\0\62\0\40\0\115\0\141\0\162\0\153\0\145\0\164\0\151\0\156\0\147\0\40\0\164\0\141\0\162\0\147\0\145\0\164\0\40\0\163\0\164\0\162\0\141\0\164\0\145\0\147\0\171\0\40\0\145\0\170\0\141\0\155\0\160\0\154\0\145\0\40\0\165\0\163\0\151\0\156\0\147\0\40\0\111\0\104\0\63\0\40\0\104\0\145\0\143\0\151\0\163\0\151\0\157\0\156\0\40\0\124\0\162\0\145\0\145\0\163\0\40\0\151\0\156\0\40\0\122\0\165\0\142\0\171)
+ /Parent 27 0 R
+ /Prev 28 0 R
+ /Next 30 0 R
+ /A 11 0 R
+>> endobj
+30 0 obj
+<<
+ /Title (\376\377\0\63\0\40\0\102\0\145\0\164\0\164\0\145\0\162\0\40\0\144\0\141\0\164\0\141\0\40\0\154\0\157\0\141\0\144\0\151\0\156\0\147)
+ /Parent 27 0 R
+ /Prev 29 0 R
+ /Next 31 0 R
+ /A 13 0 R
+>> endobj
+31 0 obj
+<<
+ /Title (\376\377\0\64\0\40\0\101\0\40\0\147\0\157\0\157\0\144\0\40\0\164\0\151\0\160\0\40\0\146\0\157\0\162\0\40\0\144\0\141\0\164\0\141\0\40\0\145\0\166\0\141\0\154\0\165\0\141\0\164\0\151\0\157\0\156)
+ /Parent 27 0 R
+ /Prev 30 0 R
+ /Next 32 0 R
+ /A 15 0 R
+>> endobj
+32 0 obj
+<<
+ /Title (\376\377\0\65\0\40\0\115\0\157\0\162\0\145\0\40\0\141\0\142\0\157\0\165\0\164\0\40\0\111\0\104\0\63\0\40\0\141\0\156\0\144\0\40\0\144\0\145\0\143\0\151\0\163\0\151\0\157\0\156\0\40\0\164\0\162\0\145\0\145\0\163)
+ /Parent 27 0 R
+ /Prev 31 0 R
+ /A 17 0 R
+>> endobj
+33 0 obj
+<< /Type /Font
+/Subtype /Type1
+/Name /F3
+/BaseFont /Helvetica-Bold
+/Encoding /WinAnsiEncoding >>
+endobj
+34 0 obj
+<< /Type /Font
+/Subtype /Type1
+/Name /F5
+/BaseFont /Times-Roman
+/Encoding /WinAnsiEncoding >>
+endobj
+35 0 obj
+<< /Type /Font
+/Subtype /Type1
+/Name /F1
+/BaseFont /Helvetica
+/Encoding /WinAnsiEncoding >>
+endobj
+36 0 obj
+<< /Type /Font
+/Subtype /Type1
+/Name /F9
+/BaseFont /Courier
+/Encoding /WinAnsiEncoding >>
+endobj
+37 0 obj
+<< /Type /Font
+/Subtype /Type1
+/Name /F2
+/BaseFont /Helvetica-Oblique
+/Encoding /WinAnsiEncoding >>
+endobj
+38 0 obj
+<< /Type /Font
+/Subtype /Type1
+/Name /F7
+/BaseFont /Times-Bold
+/Encoding /WinAnsiEncoding >>
+endobj
+1 0 obj
+<< /Type /Pages
+/Count 4
+/Kids [6 0 R 19 0 R 21 0 R 23 0 R ] >>
+endobj
+2 0 obj
+<< /Type /Catalog
+/Pages 1 0 R
+ /Outlines 27 0 R
+ /PageMode /UseOutlines
+ >>
+endobj
+3 0 obj
+<< 
+/Font << /F3 33 0 R /F5 34 0 R /F1 35 0 R /F9 36 0 R /F2 37 0 R /F7 38 0 R >> 
+/ProcSet [ /PDF /ImageC /Text ] >> 
+endobj
+9 0 obj
+<<
+/S /GoTo
+/D [19 0 R /XYZ 85.0 659.0 null]
+>>
+endobj
+11 0 obj
+<<
+/S /GoTo
+/D [19 0 R /XYZ 85.0 519.466 null]
+>>
+endobj
+13 0 obj
+<<
+/S /GoTo
+/D [21 0 R /XYZ 85.0 520.96 null]
+>>
+endobj
+15 0 obj
+<<
+/S /GoTo
+/D [21 0 R /XYZ 85.0 356.826 null]
+>>
+endobj
+17 0 obj
+<<
+/S /GoTo
+/D [23 0 R /XYZ 85.0 659.0 null]
+>>
+endobj
+27 0 obj
+<<
+ /First 28 0 R
+ /Last 32 0 R
+>> endobj
+xref
+0 39
+0000000000 65535 f 
+0000009297 00000 n 
+0000009376 00000 n 
+0000009468 00000 n 
+0000000015 00000 n 
+0000000071 00000 n 
+0000000814 00000 n 
+0000000934 00000 n 
+0000000987 00000 n 
+0000009602 00000 n 
+0000001122 00000 n 
+0000009665 00000 n 
+0000001258 00000 n 
+0000009731 00000 n 
+0000001394 00000 n 
+0000009796 00000 n 
+0000001531 00000 n 
+0000009862 00000 n 
+0000001668 00000 n 
+0000003584 00000 n 
+0000003692 00000 n 
+0000005556 00000 n 
+0000005664 00000 n 
+0000006554 00000 n 
+0000006677 00000 n 
+0000006711 00000 n 
+0000006904 00000 n 
+0000009926 00000 n 
+0000007099 00000 n 
+0000007360 00000 n 
+0000007852 00000 n 
+0000008070 00000 n 
+0000008351 00000 n 
+0000008635 00000 n 
+0000008748 00000 n 
+0000008858 00000 n 
+0000008966 00000 n 
+0000009072 00000 n 
+0000009188 00000 n 
+trailer
+<<
+/Size 39
+/Root 2 0 R
+/Info 4 0 R
+>>
+startxref
+9977
+%%EOF

data/site/build/site/en/neuralNetworks.html

@@ -0,0 +1,446 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
+<html>
+<head>
+<META http-equiv="Content-Type" content="text/html; charset=UTF-8">
+<meta content="Apache Forrest" name="Generator">
+<meta name="Forrest-version" content="0.8">
+<meta name="Forrest-skin-name" content="pelt">
+<title>Backpropagation neural networks in ruby :: ai4r</title>
+<link type="text/css" href="skin/basic.css" rel="stylesheet">
+<link media="screen" type="text/css" href="skin/screen.css" rel="stylesheet">
+<link media="print" type="text/css" href="skin/print.css" rel="stylesheet">
+<link type="text/css" href="skin/profile.css" rel="stylesheet">
+<script src="skin/getBlank.js" language="javascript" type="text/javascript"></script><script src="skin/getMenu.js" language="javascript" type="text/javascript"></script><script src="skin/fontsize.js" language="javascript" type="text/javascript"></script>
+<link rel="shortcut icon" href="">
+</head>
+<body onload="init()">
+<script type="text/javascript">ndeSetTextSize();</script>
+<div id="top">
+<!--+
+    |header
+    +-->
+<div class="header">
+<!--+
+    |start group logo
+    +-->
+<!--+
+    |end group logo
+    +-->
+<!--+
+    |start Project Logo
+    +-->
+<div class="projectlogoA1">
+<a href="http://ai4r.rubyforge.org/"><img class="logoImage" alt="ai4r" src="images/ai4r-logo.png" title="Artificial Intelligence for Ruby"></a>
+</div>
+<!--+
+    |end Project Logo
+    +-->
+<!--+
+    |start Search
+    +-->
+<div class="searchbox">
+<div class="roundtopsmall">
+<form target="_top" action="http://www.google.com/custom" method="get">
+<input value="ai4r.rubyforge.org;raa.ruby-lang.org;rubyforge.org" name="domains" type="hidden"><label style="display: none" for="sbi">Enter your search terms</label><input id="sbi" value="" maxlength="255" size="31" name="q" type="text"><label style="display: none" for="sbb">Submit search form</label><input id="sbb" value="Search" name="sa" type="submit"><span style="display:none"><input id="ss0" value="" name="sitesearch" style="display=none" type="radio"><input id="ss1" checked value="ai4r.rubyforge.org" name="sitesearch" style="display=none" type="radio"><input id="ss2" value="raa.ruby-lang.org" name="sitesearch" style="display=none" type="radio"><input id="ss3" value="rubyforge.org" name="sitesearch" style="display=none" type="radio"></span><input value="pub-2578934938711895" name="client" type="hidden"><input value="1" name="forid" type="hidden"><input value="ISO-8859-1" name="ie" type="hidden"><input value="ISO-8859-1" name="oe" type="hidden"><input value="active" name="safe" type="hidden"><input value="0000" name="flav" type="hidden"><input value="XEZZoGuMPMQqvUG_" name="sig" type="hidden"><input value="GALT:#008000;GL:1;DIV:#336699;VLC:663399;AH:center;BGC:FFFFFF;LBGC:336699;ALC:0000FF;LC:0000FF;T:000000;GFNT:0000FF;GIMP:0000FF;LH:50;LW:78;L:http://ai4r.rubyforge.org/images/ai4r-logo.png;S:http://ai4r.rubyforge.org/;FORID:1" name="cof" type="hidden"><input value="en" name="hl" type="hidden">
+</form>
+</div>
+</div>
+<!--+
+    |end search
+    +-->
+<!--+
+    |start Tabs
+    +-->
+<ul id="tabs">
+<li class="current">
+<a class="selected" href="index.html">Home</a>
+</li>
+</ul>
+<!--+
+    |end Tabs
+    +-->
+</div>
+</div>
+<div id="main">
+<div id="publishedStrip">
+<!--+
+    |start Subtabs
+    +-->
+<div id="level2tabs"></div>
+<!--+
+    |end Endtabs
+    +-->
+<script type="text/javascript"><!--
+document.write("Last Published: " + document.lastModified);
+//  --></script>
+</div>
+<!--+
+    |breadtrail
+    +-->
+<div class="breadtrail">
+
+             &nbsp;
+           </div>
+<!--+
+    |start Menu, mainarea
+    +-->
+<!--+
+    |start Menu
+    +-->
+<div id="menu">
+<div onclick="SwitchMenu('menu_selected_1.1', 'skin/')" id="menu_selected_1.1Title" class="menutitle" style="background-image: url('skin/images/chapter_open.gif');">Home</div>
+<div id="menu_selected_1.1" class="selectedmenuitemgroup" style="display: block;">
+<div class="menuitem">
+<a href="index.html" title="ai4r - Artificial Intelligence for Ruby">Index</a>
+</div>
+<div class="menuitem">
+<a href="geneticAlgorithms.html" title="Genetic Algorithms in ruby">Genetic Algorithms</a>
+</div>
+<div class="menuitem">
+<a href="machineLearning.html" title="ID3 Decision Trees in ruby">Machine Learning</a>
+</div>
+<div class="menupage">
+<div class="menupagetitle">Neural Networks</div>
+</div>
+<div class="menuitem">
+<a href="downloads.html" title="ai4r - Download Files">Downloads</a>
+</div>
+<div class="menuitem">
+<a href="svn.html" title="ai4r Subversion repository">Subversion (svn)</a>
+</div>
+<div class="menuitem">
+<a href="wholesite.pdf">ai4r project PDF</a>
+</div>
+</div>
+<div id="credit">
+<script type="text/javascript">
+google_ad_client = "pub-2578934938711895";
+google_ad_slot = "6111091580";
+google_ad_width = 125;
+google_ad_height = 125;
+</script><script src="http://pagead2.googlesyndication.com/pagead/show_ads.js" type="text/javascript"></script>
+<hr>
+</div>
+<div id="roundbottom">
+<img style="display: none" class="corner" height="15" width="15" alt="" src="skin/images/rc-b-l-15-1body-2menu-3menu.png"></div>
+<!--+
+  |alternative credits
+  +-->
+<div id="credit2"></div>
+</div>
+<!--+
+    |end Menu
+    +-->
+<!--+
+    |start content
+    +-->
+<div id="content">
+<div title="Portable Document Format" class="pdflink">
+<a class="dida" href="neuralNetworks.pdf"><img alt="PDF -icon" src="skin/images/pdfdoc.gif" class="skin"><br>
+        PDF</a>
+</div>
+<h1>Backpropagation neural networks in ruby :: ai4r</h1>
+<div id="minitoc-area">
+<ul class="minitoc">
+<li>
+<a href="#nn-Introduction">Introduction to Neural Networks in Ruby</a>
+</li>
+<li>
+<a href="#nn-HowTo">How to use Backpropagation Neural Networks in Ruby</a>
+</li>
+<li>
+<a href="#nn-example">Example using ai4r Backpropagation network in ruby</a>
+</li>
+<li>
+<a href="#nn-custom">Customizing your neural network in ai4r</a>
+</li>
+<li>
+<a href="#nn-more">More about Neural Networks and Backpropagation</a>
+</li>
+</ul>
+</div>
+    
+<a name="N1000D"></a><a name="nn-Introduction"></a>
+<h2 class="boxed">Introduction to Neural Networks in Ruby</h2>
+<div class="section">
+<p>
+        The utility of artificial neural network models lies in the fact that they can be used to infer a function from observations. This is particularly useful in applications where the complexity of the data or task makes the design of such a function by hand impractical. Neural Networks are being used in many businesses and applications. Their ability to learn by example makes them attractive in environments where the business rules are either not well defined or are hard to enumerate and define. Many people believe that Neural Networks can only solve toy problems. Give them a try, and let you decide if they are good enough to solve your needs.
+      </p>
+<p>
+In this module you will find an implementation of neural networks using the Backpropagation is a supervised learning technique (described by Paul Werbos in 1974, and further developed by David E. Rumelhart, Geoffrey E. Hinton and Ronald J. Williams in 1986)
+      </p>
+</div>
+ 
+<a name="N1001A"></a><a name="nn-HowTo"></a>
+<h2 class="boxed">How to use Backpropagation Neural Networks in Ruby</h2>
+<div class="section">
+<pre class="code">
+
+# Create the network
+  net = Backpropagation.new([4, 3, 2])  # 4 inputs
+                                        # 1 hidden layer with 3 neurons,
+                                        # 4 outputs
+  # Train the network
+  1..upto(100) do |i|
+    net.train(example[i], result[i])
+  end
+
+  # Use it: Evaluate data with the trained network
+  net.eval([12, 48, 12, 25])  # =&gt;  [34, 22]
+
+</pre>
+</div>
+  
+
+<a name="N10025"></a><a name="nn-example"></a>
+<h2 class="boxed">Example using ai4r Backpropagation network in ruby</h2>
+<div class="section">
+<p>
+Let's imagine that we have to implement a program to identify simple patterns (triangles, squares, crosses, etc). The main problem is that this program must be resistant to random noise in the image (pixels with wrong 
+values) and line noise (similar to the unwanted direct current that we usually have in a signal).</p>
+<p>
+We can take an example of each pattern to be recognized, and train a neural network to identify them.
+</p>
+<p>
+We create a network with the following architecture: 256 input neurons, 128 neurons in a hidden layer, 3 output neurons. We feed this network with 16x16 matrices (in fact will convert them to vectors of length 256). Each pixel is represented with a number from 0 (white pixel) to 10 (black pixel). The output of this network if a 3 vector of dimension 3, where ideally:</p>
+<ul>
+	
+<li>(1, 0, 0) for triangles</li>
+	
+<li>(0, 1, 0) for squares</li>
+	
+<li>(0, 0, 1) for crosses</li>
+
+</ul>
+<p>We train our backpropagation neural network using the following examples:</p>
+<table class="ForrestTable" cellspacing="1" cellpadding="4">
+
+<caption>Training patterns</caption>
+	
+<tr>
+		
+<td colspan="1" rowspan="1"><img alt="Triangule training example" src="/images/t.png"></td>
+		<td colspan="1" rowspan="1"><img alt="Square training example" src="/images/s.png"></td>
+		<td colspan="1" rowspan="1"><img alt="Cross training example" src="/images/c.png"></td>
+	
+</tr>
+
+</table>
+<p>And we repeat the training 20 times.</p>
+<p>
+The results we got when we evaluate patterns with our trained network are:
+</p>
+<ul>
+	
+<li>Evaluating the training patterns with the trained network:
+	<ol>
+		
+<li> 
+<img alt="Triangule training example" src="/images/t.png"> 
+[0.98, 0.03, 0.01] =&gt; TRIANGLE </li>
+		
+<li> 
+<img alt="Square training example" src="/images/s.png"> 
+[0.00, 0.96, 0.03] =&gt; SQUARE </li>
+		
+<li> 
+<img alt="Cross training example" src="/images/c.png"> 
+[0.00, 0.00, 0.99] =&gt; CROSS </li>
+	
+</ol>
+	
+</li>
+	
+<li>Evaluating the patterns with random noise with the trained network:
+	<ol>
+		
+<li> 
+<img alt="Triangule pattern with random noise" src="/images/t_wn.png"> [0.98, 0.01, 0.01] =&gt; TRIANGLE </li>
+		
+<li> 
+<img alt="Square pattern with random noise" src="/images/s_wn.png"> [0.00, 0.96, 0.02] =&gt; SQUARE </li>
+		
+<li> 
+<img alt="Cross pattern with random noise" src="/images/c_wn.png"> [0.00, 0.00, 0.98] =&gt; CROSS </li>
+	
+</ol>
+	
+</li>
+	
+<li>Evaluating the patterns with line noise with the trained network:
+	<ol>
+		
+<li> 
+<img alt="Triangule pattern with line noise" src="/images/t_wbn.png"> [0.62, 0.00, 0.02] =&gt; TRIANGLE </li>
+		
+<li> 
+<img alt="Square pattern with line noise" src="/images/s_wbn.png"> [0.00, 0.75, 0.01] =&gt; SQUARE </li>
+		
+<li> 
+<img alt="Cross pattern with line noise" src="/images/c_wbn.png"> [0.00, 0.00, 0.98] =&gt; CROSS </li>
+	
+</ol>
+	
+</li>
+
+</ul>
+<p>These results are very satisfactory. The network could sucessfully identify the patterns despite the noise introduced to them. In fact, one of the most popular uses of neural networks in business 
+applications is OCR (opticar character recognition).</p>
+<p>This is the source code used to elaborate this example (You can find it inside the zip file):</p>
+<pre class="code">
+
+require File.dirname(__FILE__) + '/training_patterns'
+require File.dirname(__FILE__) + '/patterns_with_noise'
+require File.dirname(__FILE__) + '/patterns_with_base_noise'
+require File.dirname(__FILE__) + '/../../lib/neural_network/backpropagation'
+
+net = NeuralNetwork::Backpropagation.new([256, 128, 3])
+
+tr_input = TRIANGLE.flatten.collect { |input| input.to_f / 10}
+sq_input = SQUARE.flatten.collect { |input| input.to_f / 10}
+cr_input = CROSS.flatten.collect { |input| input.to_f / 10}
+
+tr_with_noise = TRIANGLE_WITH_NOISE.flatten.collect { |input| input.to_f / 10}
+sq_with_noise = SQUARE_WITH_NOISE.flatten.collect { |input| input.to_f / 10}
+cr_with_noise = CROSS_WITH_NOISE.flatten.collect { |input| input.to_f / 10}
+
+tr_with_base_noise = TRIANGLE_WITH_BASE_NOISE.flatten.collect { |input| input.to_f / 10}
+sq_with_base_noise = SQUARE_WITH_BASE_NOISE.flatten.collect { |input| input.to_f / 10}
+cr_with_base_noise = CROSS_WITH_BASE_NOISE.flatten.collect { |input| input.to_f / 10}
+
+puts "Training the network, please wait."
+20.times do
+  net.train(tr_input, [1,0,0])
+  net.train(sq_input, [0,1,0])
+  net.train(cr_input, [0,0,1])
+end
+
+def result_label(result)
+  if result[0] &gt; result[1] &amp;&amp; result[0] &gt; result[2]
+    "TRIANGLE"
+  elsif result[1] &gt; result[2] 
+    "SQUARE"
+  else    
+    "CROSS"
+  end
+end
+
+puts "Training Examples"
+puts "#{net.eval(tr_input).inspect} =&gt; #{result_label(net.eval(tr_input))}"
+puts "#{net.eval(sq_input).inspect} =&gt; #{result_label(net.eval(sq_input))}"
+puts "#{net.eval(cr_input).inspect} =&gt; #{result_label(net.eval(cr_input))}"
+puts "Examples with noise"
+puts "#{net.eval(tr_with_noise).inspect} =&gt; #{result_label(net.eval(tr_with_noise))}"
+puts "#{net.eval(sq_with_noise).inspect} =&gt; #{result_label(net.eval(sq_with_noise))}"
+puts "#{net.eval(cr_with_noise).inspect} =&gt; #{result_label(net.eval(cr_with_noise))}"
+puts "Examples with base noise"
+puts "#{net.eval(tr_with_base_noise).inspect} =&gt; #{result_label(net.eval(tr_with_base_noise))}"
+puts "#{net.eval(sq_with_base_noise).inspect} =&gt; #{result_label(net.eval(sq_with_base_noise))}"
+puts "#{net.eval(cr_with_base_noise).inspect} =&gt; #{result_label(net.eval(cr_with_base_noise))}"
+
+</pre>
+</div>
+
+  
+<a name="N100C6"></a><a name="nn-custom"></a>
+<h2 class="boxed">Customizing your neural network in ai4r</h2>
+<div class="section">
+<p>Sometime for a given problem, you will have to "play around" with some parameters to 
+	get to a solution. This parameters are:</p>
+<p>
+<strong>threshold</strong>: A real number which we will call Threshold. Experiments have shown that best values for q are between 0.25 and 1. You can optionally pass this parameter to the initialization method of your network.</p>
+<p>
+<strong>lambda</strong>: The Learning Rate: a real number, usually between 0.05 and 0.25. You can optionally pass this parameter to the initialization method of your network.</p>
+<p>
+<strong>momentum</strong>: A momentum will avoid oscillations during learning, converging to a solution in less iterations.  You can optionally pass this parameter to the initialization method of your network.</p>
+<p>
+<strong>transference function</strong>: By default, f(x) = 1/(1 + e^(-x)). 
+  This function is called "Sigmoid function". You can see it like a 
+  "smoothed" version of the "Heaviside step function". It will always provide a 
+  value between 0 and 1. </p>
+<p>
+<img alt="Sigmoid function" src="images/sigmoid.png"></p>
+<p>
+  Sometimes you will have  better results with f(x) = x. You can change the transference function from the default sigmoidal function to the linear one, overriding NeuralNetwork::Neuron.f and NeuralNetwork::Neuron.f_prime (derived function
+of f):</p>
+<pre class="code">
+
+	class NeuralNetwork::Neuron
+		def self.f(x)
+			x
+		end
+		def self.f_prime(x)
+			1
+		end
+	end
+
+</pre>
+</div>
+
+  
+<a name="N100F0"></a><a name="nn-more"></a>
+<h2 class="boxed">More about Neural Networks and Backpropagation</h2>
+<div class="section">
+<ul>
+
+<li>
+	
+<a href="http://en.wikipedia.org/wiki/Artificial_neural_network">Wikipedia article on Artificial Neural Networks</a>
+
+</li>
+
+<li>
+	
+<a href="http://en.wikipedia.org/wiki/Backpropagation">Wikipedia article on Backpropagation Algorithm</a>
+
+</li>
+
+<li>
+	
+<a href="http://www.tek271.com/articles/neuralNet/IntoToNeuralNets.html">Neural Networks - An Introduction by Abdul Habra</a>
+
+</li>
+
+<li>
+
+<a href="http://galaxy.agh.edu.pl/~vlsi/AI/backp_t_en/backprop.html">A graphical explanation of the 
+backpropagation algorithm by Mariusz Bernacki and Przemysław Włodarczyk</a>
+
+</li>
+
+</ul>
+</div>
+
+  
+</div>
+<!--+
+    |end content
+    +-->
+<div class="clearboth">&nbsp;</div>
+</div>
+<div id="footer">
+<!--+
+    |start bottomstrip
+    +-->
+<div class="lastmodified">
+<script type="text/javascript"><!--
+document.write("Last Published: " + document.lastModified);
+//  --></script>
+</div>
+<div class="copyright">
+        Copyright &copy;
+         2007-2008 Sergio Fierens</div>
+<div class="host">
+<a href="http://rubyforge.org/projects/ai4r/"><img class="logoImage" alt="" src="images/rubyforge.png"></a>
+</div>
+<!--+
+    |end bottomstrip
+    +-->
+</div>
+<script src="http://www.google-analytics.com/ga.js" type="text/javascript"></script><script type="text/javascript">
+var pageTracker = _gat._getTracker("UA-3234625-1");
+pageTracker._initData();
+pageTracker._trackPageview();
+</script>
+</body>
+</html>