From 2f08ed9f1b4464dcceee73c02af51e8d447ca4ee Mon Sep 17 00:00:00 2001 From: jannisp Date: Fri, 27 Aug 2021 13:23:10 +0200 Subject: [PATCH] Start forecasting --- img/forecast-notation.png | Bin 0 -> 43359 bytes main.tex | 140 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 140 insertions(+) create mode 100644 img/forecast-notation.png diff --git a/img/forecast-notation.png b/img/forecast-notation.png new file mode 100644 index 0000000000000000000000000000000000000000..862539c58730407077b185466397c4fb9218487c GIT binary patch literal 43359 zcmd431yEP*w>G>1Q4moOkd_iqLQ+8_lon|bX`~d84(U)p8bky{q+6s*Bqb#T=}zhH zj(6?n|3BxQZ_dm&^PTg(GcPkffv|sj-+SF_UF*88b^9yIOW|LnxQHMK{v&B|Wdy-I zMi7iS(rpB;piczdN$k)>>`<61~;-p7&yx!e#L=&J zb>8uQ@Fm^&X`P4=zM9F1-2eF@a)Vk0^e1i@$nN{2KVle!D}w$+SP<7F^FNp2Vmd~D zLY%wt|Lr1@d+U>814BczM=xK#yui%NtdgTePUF5$_YR-(9fMTJCxLBDrL5lXHZuX4 zb+q){pN(20D{ZRd`gNmjYhR)m_Q4~)^us;r!vm8=BxAeNv{?G_V^+n%>1o3oH*QQ! zH-~@Xv%IcRY7u83k=+>1tYCF|>@-ql zdjZKaAHG}ks{b;qK;xyt2RS;G0g}NK7|3TnON%LEKR?;Jj*crB7#P^M&&FKV$Zv!s zB+v>vt|utw6Eu?$xpUuwr$0VDT04mBH0D*ydyXL)M210kZg*waSLE!Ni2D9(UitUf zDXLj=)R=@}%|71V&kwg3(k^4EtE;D7ev>iqoW`AHx9#_DF0%nn*Nq0!uA*`Gqu<`> zDQxdJ6;EH^3&a;C!b;xw`Ay5ED$@xFP+Y!zd3|&9ZCRO9&I~UFzQxG*H}&=6 zYMPpLb#+)FAt90>kurEC>recZGL;z@cK$Yn+P3wZ77ygTOPH#XiRCo>F{T87Tg&kHU|KYH}@R~Wr25BwLLBSl{0Ej)|c>f8{#c zsKWb|S|2KA-j3soO$vczAfbV1$g1e5`=p z(e8>sH#-|!iu)H66O(}oC!W`6etLpTooAsVU$mShwZq@0}v5LyO?WIs)LaTRAl_hbGh;X8cf zB(g7;boK8{WW*2qrNy=*xRX@F zYMn~wfpP~9Wo6~}(b4|!!o7WcA2Zy^P+@`JWotT=3Zc5ic$6#|Tnp=$pq%+WEbPsz z{+vYkgLtN-+_*%>`9Pz+zn^rjJ+>}QE+)l9$a%Y_%y!Odb>tqeNtd`nJVHoV*|PDs z$r`@H&(Dt_uCA^rCmW^X?yvHktItkdQqt4^R_=|}txwi<<{L<<eVZgs=`7ZjUwYW6BEzTkJ?)uts5+Sxd9nV zB|{Mdq4ln545q|D;1=pbMNV(BTfTVl&SA{yLe{g~#?f*I>-C9;Uj&`5=I7@_1d9(A z7%63`vfqq2JDZ!HntH2IY)U?hMMuxl2>J5;=g%4m1-7$oZ&FiJb(Z$__AVfqYp%P( zLw~M|xCfn{92qQhrFQWaYp{&MCxHFIql{ZV-3%YUcY{A{3ihee&oVLgVFf1w{EKPqbnXx_pCjlMH6eh zafz%Bw@oTec8Bl%AR#2|40K<4*%8OCI0$KTNobOfgq4*wwvm3~j7H(|bwNMdxpoXh zz-~Sve7U-9Z0tJEi?#M?vPt%Pkm1J24|_b9y;X^`)02(fXSzT8`=7l$INVuEc0W19 z4vF&flkpuYepNR%Cfx5-?ywfA;{ee$zQb!dMg?p3tF5gb-v2fQLxhg(ex@4hle9}; zRkH)G2=5~hpEs=vVj+%W_$;Spf5&`xKM}aC{lgnpcA_PMRhc|nHLFloO^pn4F&3g$ z<$5oYLpP*!%o#73l25Ym#qU4g)Oku4i3lreA*XSfbfsWM6c-e5Tk$qN$<`sFIaqTyH`O+6QidK~?xkh#*-m0?oYl(V0p zBA0RU@kKqQ*Ht_SPK^Kf@kV94dz|b;<6@1Jl$5-D{{X1t(I2CuqviQx`ipI5WVYAF zY2beay3NhZCYF{01?(3S^y=OXRJ!oN!^X`yqlFrBwBE_#&cF|Q))1rlRs z=R-v%=wiXYyjY%oogR22;+Yw zfovJu>DpsjyzM;t!`{l?-Uq4&uh~nB^OwjFO-)UNL;kYP?q)MK$Iq*Tm2t~%Y{&>C z-42_uu&{$1W|#Oc{ikZWJH+)(G#+e~rzd;kTkI|?Lw7!M=w1hO(d{bNFMZ{sQ(41#&vv!0 z!-^hRah1=GPEL<0%j#x;ZNq3rH zEvdNGwfG2O3mYa_|84AAdgyhYj)_U~rMsOSF95OZt!*(x zpxFc7YRb+RA(v}(baec-vq3%FH@pFLo`bi*z|P~i{ycW4?E81=-2PnM(A39K`mK@d z2XWp6gkd>RdxZeK8n+j^!#;h&$yN=7wDlZ9e|{i8Ssu^LqF}Fiu=N<0W`F$bG=0zc zVr^H~RXoz$=hmkjyE30J%YNY}481P&p+)P59WLC)ht)e-Dw$U$L$7O3|L{kUx4(X^ zj#qOaI7pIA1dD`@uCnrn_4W1leI_qnFhOo$5#FD`LZOYxkKEi`z?X5`?p-O51eW{* z1BoJ7pGD4M5vZ!FCh0c>)YjIDzC6s+tFsB78Fd)%y>;P%Y&3W3uVHC%l`rgSc@Ku> zg2{Q(el0C^r9KXE%M?xgGstH-UZw3^ZP@&QYKctP)iJ~#EsnW6R3h#bio-&dV}g1X z7CE*R&Rf6vB+6`N?DPQ)5y6(@81j|gzKY_Kx#?-|WU0_3?ef>0JUm`d<#(3)tv9D1 z0Vb{)9i=dsKA@?tF)Qc@^X;lIhD!YvNe zXbA!kX8*4BcD(zi8iTHaTEmu6>YA!fciti|=L`~qMN*`f*HydvPA%HGE* z1aBQkKe!4yZ{t8gk+8AhUS3|FSXhW&>%Ow}0dn%=sN0k6vG*NMkM99CQ_9x3qw-BH z|3T|`r3>lV=@G}b?Cg!*VauiMHD>yxw*adl$}U~KdiA-vIX(>y4P@0+#U*KZ`9=Wr ziHbKTCMW6c9L|@=)jDrZ;u&I&V&U0$8JWNs2dZWdA0HN>}xm z-Vr{SCR!dYU5hX(w_AAYv^D#)p`if>K27tVUR3(-M;S(dRCzCcQ+zKgo6#&O+FpB! z2kmJy1^k2zN3lk5=96^&k?+>IY)jBM0H$N$< zvS}tu;va7`(X8+7@rfEXD$RR~+`fIgqdY!y+uF*?Gc+_5?M++4eSM`uOiC?AxgXvT zH#E(HRH67vgf{o|#7BM~kXT@jb~y#hb+kK9(l@B}#wEM|n(d=ZYNnWzd4&le71&Kz zY{;tP|J6z>2^Tc2#q(Py!QZnp?hTzod$};eo(v_dKvE9L=g$++zcBk(yWQdgolW02 zbuXTm?H>mWX*k#OzO1gU`h|uzIgGpc0+^!w`dL32wh~(EFtM<#x963f_!?NHuAD{v zdhdRA;;_Bgi-$lz@-)fvGAoXog{6iD1-$C5FS=Bc3*`4~ZrrsT3-a_t%!Z1ZGwfq( zp(1jc_S{5jAh{Twxus1p-@J+Aur}5(@zaaTW?I_BSx<5yno|$6>h$o{z`y_*i)`{F z$9n)|yp)tNqBQSEM@R9AWUrf(LXt&a45CGWo#h%Q9!|Y8`X>|V-09^6UtvbB8HP_)`H27gmK?KkQ%1TJ97Px%t z_;RoC!Aw6aO$hbiJKY^u{v)6v>CN8j`28Cn(CY^yk(1-R$5+tz(=1u(ykiqv z1+fRo_Tvvjn&*>)2cf`arh0=&+^W76{x1Haku51Ba}lX& zYr7;bFYg@mDPw)0Rw1=El3jCRY6_E-loT>L+js#qTgrow^{3S>ptVvi0}Q@!-;lnG zh)z62B*_f&a&8fagSGWTNlD2#qobWA=7<;}m)CJK@?ZGrBDOX*xWosF#Y$&;m9HwzJ&9xzCP!zC+$pdHP?(84 zhtHfczLZ;JZfSB~_+!bLE{pr8p#|g#EIbXO$=Bb*kv2aI~Mg6vJRS>+k>3h1FD*7m` zH;|XeE0@)qWY84y3%=~heMH|JY|Q~VCAPK^>G&2B!)8Q81mo+qG4?UnEjf$_hcWjZ zEUH?e-;?gid?Ll7TWtC?-)=z}<#!k)bnHTDM6N+LzY5%zz_74E6l4qrst~WgAm_{WFIQuhu^yoO!ubVKC-iDeN7(h=XLo>;m>9@6aN2!!4yB2p zoud~P9`g%>T9^1KIvQ;Ofw>wk<>)BzI6YuyQiuzhX^k@Ylkl(+rz3%loxL+nj&k8- ztkPwRt=#&rBmit|R@HAcP=$gCuSYzG6@F7*e&3T6*Uiw-FwV#q`ck_gG5xqd z9j+kW5I~Y@K0jXVkuw978!O;pYy@&9UC4DI(`i3{S+DAdk>(TIM+yq0P<^{zYk#+T z`0^AWwFEF#4FyK6vc!p#zzX1`D&I19F*Eztehfq=4iIDc`VAL}B%~R8VV|b1Jq5DC zwhaX;uxi8*8C(zT7{^~R9@~sZlh*HV{7qf47%s`?b!+>?RSTOC8fpnh(mIzYc%_3V z?#RGLeBvCcy(I@rD)AGRRHkmOtH|U@Y^P~~`EZGxn!0*3_h830GP0ZIM-Y4|3&W+B zq)?YgpdaVcvr|z9%FLhJs6~1_8@}^+tLkX=KExO;v^cV4dd<5UxE>}0(bY1hreAM9 z%lU%P0`)4cq49a3)G{961u5`)Job*xjC0|X^lnp@j1D%(TC;7@4CdV^#q3Rh_MS-) zRt56AJ_P#E>UEU>$kV`~tB2Tl+)xs(cgM6x`q1KcG>;T~6Q~{^FE1~Cx5Fa;WY)p$ zu2iXz`&(~~VGW8#U-f+>gza1)c*ORf&pW+erd4JmF?6P>rshY&s*3aYajSCJdYrJE z#?WgZ1@eK0L0??tZEe%VKN;RJ_iWR)M@_p@{K>ct{qjmnNz>(Hx$TV^<8-dN-&<|N zLw8>nY4^~OMi7n2SPvb!tjWDHIu@2wRU77}j#TLLy6)a04p0tT#@m=$bo04LVprnU z{DCp;a-iV^bfN-6qRT5Q4NwRH=d;ZleY_6+;6i5;P|o{6Hgf@z0T765S+Vg>`0LF` ziaVw^DsO9+uvHz*QW8lTsBu`ns&+s5ldZ{j?%cUgob*Xk{%EXi*qZcxi{AbLd4Ik? zHy$^c>pG-nfRQq5_w5(G-(X$%%t`+NPghqL3hpCk5lFNZfEbv9q=mQdvZYfw%;ru3 z!4muD1~ao*s&X_xeck^K8d${dKYqyMD#QvH)k+8DLigRZ1axCG=hY8HKoYm{y=g0dqB|IE{KvGOW z^miTA`Cf_Pq^M(MM+3Y8%^90zZ0ZH)bX_;FRlljdft(}Vz2&U-{>$<4RtFicF@gKZ z&a<9eUER;^^+5rm$(;$12ym`kX$oVILfcma~Zq5D{D>m&#`C=5^ zB2b_MdwNueq|!W6SEk?1+i#S9xL@Xa8(u|&A0b1YX)lZ2QXf8Icl8w3g+WMYRArL9 z2kI%RTz%82q)1FmL=wW+r6eUqvsGhQ1X&WH$LkdqxFEWv5O3v7;*K~W)<2s-L6HKfdJR$=Aw_Uw zP3P5Uc_k0l#_i=n9_yJFlH{JJAB3YS+P>U-ax+3EAJ|cGQo=QSF=1z$C#M3;b=1WPK*z z7Q@$JJgi72aj)!=jLg;FF?{`lsvaI5)=Pct20jXgyobg(obRA72n4wW4`kPkoqk=) zimDHrWz#Q#f+swY8{{=*o8Q8P(tx3S^?3A z#newl#_+|9fD-fJ0AL5j>g(&LHPS%iV4M=yNbd9lVuV*3e?*>J0zkwC|fkSv=s9dZ?W-$O3bm9K88~UOEOUKS+Ft5exe1j;ihV?4tqsK#W@6oJF}h@TJEBx zyXVMSK@0k)Dvc*4?v*pn*fw_nZt;P)4TNxeVH;HI6muI}Th`alxGl#jZk@kObyHMS z^kVS{U_;%q6_2x%(xR2?*RM<9_o5|69C&_l;O~6`0|TLjB(b)(PP?4d)AH*V0g#JV z(Iol(^@>K^jE->F?U1AMbUM?;gZMEGC*GyfxE@NcXB&mrD?&g+k{+H~US`9@#H>fl z4)k9+Pvnz3ah^SUwokbj^dDY;%y04W@fm9QN#9)$3gpW`81aK9EC5b0pV}ZKEa)%9STt-c6F!g;^V8!xZuMD}i0(8T% zwY4=F27s*7^L}R$B-JD+@ANlrhy$DO_`OnrffhK6IzZBHBH!SYDiR>aKN z$)0ZP?Vr`hn}LK$Y!!eI@j&_xi-<7#e?Hv|RNN^@4cDNGI~xtcdkO%8tg25v1Ku>W z4^tqen z6<#kFXk?g8y2M@=uC(vHXoq|b1jr^;_u0W54aNf~>SWx4havz6ruua~NC1vURWU=u z<33|Ec$67s+X3j_tBI17l$lBHvr0k%MX)6Q{riCcEh+iGCb6p2TKdJ7C0-U}lmjdx z4n)=?1J3Q{-+(41AR&8eWZ#zgB+it+HZG&0@{x?s;)^k~AppL=ez%%@TlB5Ey4uRd zrfX?NUu?x;=b+4(mm(k3L$2pF2qAz(=k0kqZ$~b&2T+!Ml9Q8B;5DWX9Yij=ZRFo# zJOH1AAH3E85Duo*UO*2+59+*{hQ_SgWlzVY>{4+kYIr?Q7ocktz!z?AZl=F??;xUe zaCn$B;|tgx;&43x-)xTTf<0iydB$m}FPol|QxPsZ{bS|nR~15b07L8ZEF6|Ds;ZHW z5M8WMAObtBgvCLd#Q>7>hDNLMPXNAtkW1A->Ym7gFZ6YiXs2GA-_WJUmaleSB~cedr3NetKP)3Re&&M1$H91PE+TeGhUW5!eU% zAWzfN(*Coykf8Bk=b)LB_@WQm6+C+M=paXo2DduXY#{Gpkx92C`nb2Aq~Xwcs}&iC zp|~7KuqkK+4{o3R!1Vyl1)+u9w4q`BWaIVUMto3DZ=fWefJ@HL)@b~uzim49Z1L*SskQcFj{`|SWH|`-v zO`((znEXvq5igh$7;fA+2g%GAoGjOFYxzOd!$2TCy_mWg+)-PL373I=j1n18nml21 zCqcUo-MeRcIy1`u*14KPJA*3NsP6@m1njWq5H<5%sd%Wo3O%8q(`J2-9)&EZDG=j+ zA3qWoS&ULZ!IyxXj;eVXy{7dI4df|fQ0wc4hbbXUIDZ{B7YPapZGiNIN+qa_iJA%g ziJ0{#Ykj~Qavr=piKRbYFAu!AB63Us)E-JPC_i}u2|x?{9R_d{aOcihSs5~`amq}9 zmIDB_{e`tlyWBvlrKQCQARYQTEa+lUsAw@(;SZ{y9?*1YmqRlKez|{6iI1nz*VpHC z-Zn;1tvf;S=AR5DhDnH4inRo0BQn4@I0&qWIL>l zjSVt>tAxxtR&IduPtxV-V8MX|Z-i%11m=ZJ>pMYCP7bJaoc4a_?Lp~sbab46e6qf? zBN5Idm-qZ9CQ{=|i~1cC#7fapkwTgWkFAhvb!rYghT%abvlCIPW63Sujur6AY=dg~KE5l?UkOvO5k z{RBFRJ!0igA{MBixrG%nsiNI!a(XXz&+fB=$IAe+s#D)RApTLZ8n%P;7zK+w09`$h z?7=z~M8<`A?bix<11uR{!z^SqE=gQ{^E z$#{}}S@>|CVH-f=0MijdjqXmio-X zTiHC;XfxB22>LXlqjezn=uxw^D39Br9b9-(7Vq-qMrca(UcO|OW0fDp*|gnX*F#Bl zSk>^HKGcW=1QYaX@8t)RDB zIw44&o0#~cj+BN5iO(-AEJ9EcVdl%1FE-nE{rvr#i%H*ld5PuN!~Ijj>NLWZ`2&6* zEOVt!u1P)&Fq-}Oqv#6Y0{aJzs66pSY*hV(BJt0nPzl)KK+_@y!L|Xmt;gsiWBf{H zJhC@ru7jEpm6n>8HU;)6W9Y=?lb$KBfJ3UJqy&_xZ@lb+f>iM8jE9Oc<<)6u;;C9= zVq)sX$3>*Lm9H9}oyo@us>}2+fLRIl!Ta|Fu?456r}rW|s%(s2y*g_7ae8`Wi8`@B zYwr=8R{0|2eoO`}h7@RM8-p+7A9XuGo`I72v$Ip_bEKQ{x-M82%H%sD-Smfm!#f64 z_uY2x^6uJr?l^vDq3CGQV~KR-%s=4tx^!Jo&GdJ3nrar8>%J8fd?JaLj^NIGU}nY! zU9bM4qF+Gp^5eQQ^JZv?m{D4{VsFfa-{Z7mY>7m4-YxcUbH0-SjS)cF-hh#K-Nvij z!2Kf*$YFmW4R{F9j;SEyUHS+@#AU$|1k?q~l-gN!&b<#Pk&?-O(36M@9EPetX_ovFu zWg!y=wOr~cg?%rY)jRuLJkG7q25?6TrRL+W%Z6JsF`;YY z)uEq0k@1=h5TKGHe1$CmeeEro+}ZkLdj_Hy<(L6g=JE&upZf6O!+~lK5pch1Ogn>R z%vad$u;%Aa46vfIqudZw;q#Cw-&9wNfR`yi>jDYK{nyJ6fPeY}@83kj(C~&QDc@TT zor>?jocds^pc2)BVKI;{h$m`Afe!p9+F%yBdvWf3PXP3P4G{+@CQBXgb7qIuKk5v0 zB)vt>9eS5#E2OrmNs1SkIktRpy5Qd0Tg+^y`fKiW0t zGHM|Pv)nBNr8uGZFv-UYXGiTUJwY+4a%LM>qbms2=a5e58T#Glm90C!K4qPpnMvhg z2Y(Mj0Gk2c;9neAJnrzW`(%v@{qSd zb!W3NmI1(>L8`iRi5RFV{i%BYT+)g+x4b#!x}8NuF#*;G>w27^IXADQq!Gv$SmhYk zBRQpJw6H+E0mzG5T!4Gvv6mKSl#!La1YTOZ!!MDq2N{fezNvSE>eUAD9+pL&CEfrQ z?F#UbvA5;LB|t&(LAx(!4!^1!jEr=QjHv@5Jq88;lJ5L2fYnH}w5^SsUNgJ{>Bo>i#zQ^Cvig8& zoSWvtdHONw4FW<&e*Oe_QIeUgQnk5YaCO-=J1(=s)jaU`%cw6Wwx^g2p_Tw8dXs%l zIT3S}{1y{^5Ic0`}dBF3rQaV?~NjkslG|pSEw-+l$z^C!2(AwbJuV0+c?ZCgLeQur6^V>JJ z?$UKEjvt^yS5j&V*w-dRvJZ5@)I!79Y4w?Nwjm7yHVD1Rskd(D1A_gGbCmrzM2?^L z_#V&1!0&xH^})NA;F2CfHc)X(es$b3=sU#KMTulpgU)@LJpDyOaVaTM0njDxx(P60 zEw?I&J3A28@Z7&2L?twou~Wr8*>M*np0@ZSqFQHDDyG3vN*PHvq`+)o0GgY0FP&bUNLRvasFcZz4fiVrd{*liPhDd zegU+r;ix{oiicF3?qJfhtn?s_$glP7`h%=z!9xxrnu!|wy4NTu0)Q&8TN}Zx8}CJYXlGT<@~K}CgO*^#UmA3)89d3PT{nX+ zuPz;2*^WHN>NPVzqj_cCHm%<0UCd-;Sx+l;Ps;SUHAm( zR2rrp927twyS>nhM|symoCDFuSUXxw0-qtbu(CN6T|Iz!*XmHdW zm@*%^n9_;ORMJvWw~-(Cu9%n;GfNjSqedD_{{tihKt33_657F`DWztTXYa*&0^GQVt{KKli<>X`I8G0;*#m-39z4s zr!ppS^KhN|(yWi}jgyw2eCi}RJ(Gd3fIb)Xi5;))TT=--N(v%If7@7BvumZw>p(-R`sZ`^ng5gDNQEO%-loLK9g?Rg|Y<)+o{ z{@MHO^+bP$8_sUj0`{1QN|q}ASnpR11p1>3h?TYVt&$6(ZojIWg5539FI< z(ceLrkO$gN&Of_YfuF#JkP?p*vET6&ulOFlmF91dSX+iUf$RAaLO9~{MPxl&gM|8r z-OqIJi!VTXq-;jy&f*U83$*j$OBKL1(=sw*BkOx5=J(`V7SCO%D}PO0@w9bE4>lM$UDwyRXcDoGZCH|ReU0s`0HdWZjs)9B(CBU(QS{~d3Qx~YE6mm032rxZ+ln@-G#2PYpJn?pM6ZZK{b+(U*4Bp_dAPgeP>Gw zsHQ(Zp0cj?`1=KLfgh4F@S5Iyo?eEziGN<3EdRX8=pKgmtq8xTSzI2gTCQA^`BH zWp#|&<8&uVC+>5_l^R3D>u9$jsLXcI$TxtU=~(3(j4mz1Xi{=h(9fC2QGX2faf;UX zYHso!>jVe{W9yAi&i0aBI@o;dDx`l-bJO3?@HrS7<_F7oOa)!uSt{*UotRe4SJ~FL zUk+;uiSSd>$gl&<$tVfFK}nA-{JRzd&;Qhm-0$Q_=4}FZ^54Iq!unbGgkfzdsi}!G zQZ$q5D#R}mCLUE&)bK?{em8&V`Hv61pCQga}`|DkKg@E6E2Gq*zg!tf#S4UTc!4Xy&7~1Rd+3-n#7s9 zG^dfaX!Xd^af=D=+y{g{Td-rR1b`BG3_bEyD(Zaxc$+|s8~pe5ntpl-$}9TAyr&Gx z5@B!6{RL`y-?mdUXmu#HM%(qaM6Bf5tlY*-!b5(KRW{3CjPxw20Iu_lj$WJn4Ai@O zw#4Mgj_gtk{<4Tuq0fJOr5XBS`s3r1jmT$DWi`tL#_0r~bsXVlxqtRrU$-hU?yU7w zx=N^IeJ4d%WLYC0v$~2`4nIM}^O~;f8|=WX^79W=T)|?j4%45;BSqwKFJ8<>Z^w^R z)96)-G{Mvw(4fX|Zvp`3SI+8n>~6-ziW3TNoFBjNJhM`0RmizMM^Hr{y^jFp%=)TI zm*+1{-jFlAoB8B%YmqnHjh?UdEYaT6pe_xjHn6fQ35LND#N{T5fGc`Sy61wrCN zz;TQHVqw{Ak-%^JU4N$Z?w8mvoW{Qhp7=uV!sr*6#VTwvI@F^_uz+njl&btNTeXme ziI9+lER^yF^MCe#ZvV5P;eXtf|8?^CmK?Pw6>vjZ<*)HkZU|-uz*{`Ix*7spd~I9M zoC^Ajh5~JpfwU|v7r|(dp_qsXazH2W(M6Wy*8x&ULE%uUa&_8X9i@cj&eiG%x%m03 zS9s9(x`Ith(l72J!>wD74$n{*(wnG2fB!eJvDE120L$3Z(<9m+3&8`#0t_{y%{5R= zY}*2xpx_Xil!L_hwRGpIQYWb6($&9fPoI8M>+G5A_5jxDXAfP)uJy7JObI)=uL0%s z+|u$2D*o>8Cjg;{4h5rx>>RsO@zlB#DWB}Vu9&GQ3s@EcK#-sT4XO^Ev}KR_@Zl{G zF8E-|sP)9a+LYS`4hm}4djP&LB|^z!=PW0Rvv9@?6Qwn+PA)EWO-&cT0|7D>AvnCX z=6}HGPcU#vkQ^rqJA}PbQ;ib)!=j?99S#MHR_>mCH}Z%fZ<>HCg8ioYg}|M}F8Y7a@Jhe&WP{9pXHe^d+EsPJ1Npu>##F3~I1z z6#X$Nw(Q9u6h2z!L9OkeL4aNK7RVL8;4A+N^CASI&p~>Exu#l>=79k=t7WG(0H&A* zP$)K{R%}Wle0IDg%6U%6{s9*qYN;k=@6mqlJ+4^FmlrX+d-D>=|aP`Q%vBC>OmSq5Kv2i?;>r9;?PYn zs;sQ!f*~bX3`_*#JplOipJQXE5v@$j%zuIBLXaz0t}N6i(t^xn4Wnx*X=$~corGZe z2N6)~P1FTXFgwu!XoeY8RMUY;bJ4_`Jp$eYejy?6z-4vt*Ga#Ki1sDuR?Ah2-nz-bHS3_NOE zNKR9K|89VV2xe4BOwV#cPyV7YN=kYG@E9mkb=}=ha`#j)4iCOvvZr#<2AZ}JOmIY=Vbr(9XQ7b8j#5mJsJV=v`rbj2)Kr_oN4gom) zXbCdOoAwG21HdT>lNrDTG6Ru|3X#0p;QoN0@dtGZnI(Y+MyJdP ziHO22slf#>1wNB&a3;d9{8;!W0qR?VA?kx1uZ)!MjvGnaC#Mr$SGEhh{i}{dkg0Ad31QK;WEQS&21p z0;eWwh=ysuW~$aon9D{VPJroFgv`R$5$tT+kNr(Aje`CTw2sHZx{x@QL63UTcI&KZ z%*|KAdgTnu1B;$kE-wL-W%o7G{`y9QvC=zQMeNfez1>kgNysaLaugs8sBelpV{qNjupH*-Ns&N#B?`k-RyGh zfXNW&C=QJx={d5J!onxVmIek4eYv{14kxIi`#e6y&9330p^Fe;4K5KZPp^W$%&noC z6!j((>lDUqQ6I=ncMj0v1a@GqD+Z=AKeiu&`#oK@5X>A(Fk=@hFbe_2eFCOgZ}b2P z%e$6BoFHB%Dras--s9p@f_Pr)S#$E(15iT>Zo?k+>hC+rKt#YwE2R{cl=PYCf_Yv7 zT<-eXZYw#nmJl%kMQH=iQhP&4?o5!1`aykdwg)FONo!Ykw*<&6cDql<)-<|b!zij& zzPsM8gLP%P2y#{-7Ki2zQZY%W-{RCBa4 z`Hyr9Uj);Kg7@Z^R|$LsY8EpzGQy?vhO<8iDesyQ!0Oq!xjnlGR=8;!m{`uY92X9Y zjb-;Hpuc-p6}4IF+{@6Z{4wU1h`XS*;qs|W3xpr#x zRk*zcxk8sc3w0(aTKUd9rW>FGLb{g#FFDltZ;nSRrAD72TL*%iB5^&HaQ#*MJdBn~ zeA6mZh6002XS25f0eS;krE~9x3$Id8d<|~`hq%#oDCfk$5sCVZTu%hQ)W&N0TFXq@ zB0eL_)@#Rla5tEhuj?lW)Lfg2nT9)Chlm^gYHg& zj!l36zUK7wv>z-L$_uy5xE*dQeR~yfXbrO>Hf11}g8MisegklwerJ+ILH`isC~Ny^ zFjY@M6H@To0?g=W4n+TatClU!l_{iVm9G(ji1dm_hk-w|>Uf;kqbFhv%EJjJYT)RT zvaTHhpH0Id_UL#T|7{)BOZ#}Vw{Oe4AHWcrL_0IwlXc^D(CMOP@b|}^=BO)6N59;& zm5kd^%~F*+Un+`&y1n{4Tz0b(sq2HuxJ2`1%b6>I+@%Kr*77>c0z@sqDLko#T2Guz z+&>OQtrmHWfR@;72W!z+o(-?e$%Mnw$0A_U9fp!@q7$>Oqq^PHroIxZ44pSjQm9nL ztK%45sfUuYlV$=gv`ZgywaWB4IaYMSVeM{3M%dNBFfoS$m~L~}AnbN>JPtOEoKUZ* z6s+hkBXG6q;G;_5*iJ9Z3$c7jTp_4rE5dT;^HEVVT1bu0pqa~0Y8bB6CQlR`q!2W8*6Gy*J6sOx?yucCVJZO>?5Cnt3EDo-XOs3>E{(SYs^!yn6%ob z140=0*jnkSsznl!Oi;F`_o^?{&yVwa z6&!65@eP2v7p@16Z!6z^cKoGAbn)|P1T~BuU(eoovm!)K>#C8(oExfI$Jz-}$ zZ#2muHkL6i>l1H&nt$yz?JP0ZjP3^)>kKzMI=B~}Q>JQLl4bpcArYXYxXs7rJ{Gy` z1&^kcf@>7x0Su2Tp`LzGobKMx&u&Vfs(E3Y!O<`cP&euU3`@w$2DRH5U}R+OJ~b&Z zXu#cn_DzkhpZc!ZC)YFZpP!|FcI{q8!&8=ja2#gU?1+T;L$p?4MR`4B=-_rGms0(+tMM8+8jW0 zh62}EUJ(2d8|Ju$@#^YFt$;iNp)JEgj*pLngd_er=t@hK4ET$WM);R6_Z>xpO*jPX z1V|0A7TY1u4PLzp1n&-v*HRv|k7WsUYMO*9RlrE@fiukUNp|lp^(#~f16Fd)CeOq~ z(DP$(iI_k;6n%LDw}7UV+JuK*0lb(vVfO;S`UH+LBmvq1RM7s^2gW;q$Ps7gw5X*3 z>{2iDkGqySJXHB~VC0QV9x*EaL5(Bl?EukO6l^{$-^zJ#7AG*hVqkO#_mIIvD7ou(I0of zyzmZk?{gT_lia!~@3LD^6vOCKXLHSSS1}UmOhXGR6zQtFDtd?e`((ex$~U{?Huz|1 z5irkyWJe5)yB7=+{2UshfCfVfJUyL|8-NR#r*2$^-CRK5+TPv>Ct(FgK78`ixQ`l~ zoARGO-+49b@|M?PgdZ%MXmg`kVx}bPOATkmz;%&d*%?K?B0LgKFjZrD5SM=*8$F;b z1j-k~{rkhlKi-1ZmP#0o52}nOM}W6x)FplKLK_@M=nxFpl%N3vE$Qjy*u;@dni)U(ee8jFZgtLzgO*&ErBzp&&4L^H}p0z zMF=$0IZz}3x}Xgi_^_}L@V^<^x&8iAad1oKWPj3^Mjt{7CO>PS7wAlVOv-=nULxEh z%t?^KAmfvHTHTmiyHC}N4I~n&&tQvo8V>7a!bg)-bXPJd#`*mxRkM zeF?lfz@w1VFMxgsMmuR~=?5?#4^0MMYzK5^(^r1OX-05_Sbxmz>ipf6l@&b$121T6 zz_%%0WeI@}-N*BKKSK0?4bKstZs2v{6`}l!7{E;Ep*&FoCLA>JamjRU3kV$p@MB}2 zYC$yj9=ShsRv!nKE;^J1kx&PjEeFTFm^qe$=S7gd(36jQu-P{&k7re&g{FeGY6EII zre)K6m>1Fq{Ts@EErbSf{5`PX0qDn8%!{^vqO6<;gHB-A#|GeiyD&g29>m+h+(GOf>;9@i`dG-9J1;UFc+Y&B&UjnG5yxKwB+i;f2#R zrkCEHYaB8uT2eP}TztJ^zj?p8A6#4D$vp>+kRVLhR_XpZu-Lw3aWVjhBP|aVWO<-tLQhKhWBumeugV$#*}>50sP;=I@O)LlVPF6uE+{D}onL!< zGl#Tnl4Ec!Q-t54x@U6^PGkF-TYVY;w?cBXd!ODm zLA!$_#k}*Bol?f~n??WWC|HmyX(f6FOA<9jF4n2+7=W{pkc7ksu7uV@emI>lDO~_u zZa{k7`d{t6c{J8*|2BM~q*6)9P#KC0WlAXt2@U2bWF{fWCX`u`M2e81Br};(BpFMT zAsI3z%9J6IS*GVWyYKt=zUvwPeAjx{v)BG(?*`X(eXsNU49Dj zkf;Wr85fgd%sxYX8BXRsQ z8r}d;HWL#r2-s`PU~y=X;*tK)*{+lusqjH|;KZIg&R>flTCOOrU!7EJcuXfaNX%); z3b6xa4p`uF@LoZI4QIDE5I7S?G{7NLujCEV##%SMC#mO4&RZhO?9a>tlaoj#3_wIY zbZ7fv7C>l-SLBSqJCPo(6gXX4_KjDew&#h6%l-0k>jut_y~L-fs%r|3qR_8mA%7&g zD)0BWNC-0}Z*t*0r8L>t<=o{qgQh+HEv+{YBGbjW-;fvx5rv6c4I0q?W2L117=qhI z=-3ZnZW>c>nQq;?w+X-|6$NRCibBNBa=(}<9-G{vt0=_1&uF#y)U&Kgq#YnsYcO=4Jq|)J>iL`KS_B36QWQ(v(_`)ptkQ~o z^k`f0bzi7}JKqzZtY5O>r;=v5uLQxtNAVD#2co`SMN3QQ!lWsv0fq2_qrO7+y3sGx zAnwN6xsFQU+KIHxtUE0^vR-cFbK1_4*u`HqYaBSOw#&F5B;kgVoXNRQT%5@wH#b)S z7gQ8#|7}NHnF#U;V+{(?R z6dGE3IiN$1S>sGUoD&(P6&TWB5pSKR)t*kU2858#7co_Xt&90gj9BEZsO(8nKi(& z_ZZ5E)zB9r=bJ&4F|mZSUIyC~AiGL#hJ0#|_-iixEZpT9o_neN^vnb#n!CGq7^zyW z5B@@PPB*$+I@oQ0XC(SvNP^b_b(4egx^Dsh5$vNw{JTJIq@5(vSQ033k=ofV1pE=h zbO$gfELF+MZwQ+FBJ@$uz!-Vh)Ti}t(BpHZUqnJ_{69tl0QS;5R~(srT# z@Q{$WYu5~o-7Z}6icl(?=P@rcrbAmCv6Ge*2&l#0qfrRTx&m1u85tpH(gf?1inw)9 zO7e1N8E)&e@B$eA{Yy4@4>-0FH5fnnVQxWA^-S1E!q?CzU&}?IXQ33xhM`8OxwUgr z3SNVV<8<7sV6Zt}$zjL3B)$6Ym%n2aUW9U&^}9!PQE;u|X2q_xjCMFwF`gFe$TBSi)Xs)+8nfwP-Q@5{jtD}-+l z@`y#=8aPbt4G$%@JvNZk6s_JZCbrq|wPQ#PQBrdD=)ZPMdn2R_3G8t|9tOAk9o!Q4 z&+?=sC7D79PkT5Fn2vzmNozr^lu@I9=*Ks~ zf{HiX8u{QI!G91T&PO95pZoE`imyxXr{nlo(d1n-PBg9ER<2rMl4vc&w0ep5g2Ytc z4>7(3;B#Qy^U>IzudCVy5)Eo^s7}J>IimG+zTmi@cU`k)P0rOZqlyWL9|HkwtOZ?8 zu^j`rtP>I4edlnaLu-(Xg-{0xJb(UN!NK7<3M^6}J*ThV_xk~ZJSxgF1DeF{1nByq zs7Oq55LLF2;B1W!Fg=u%?R^uN1O%ezsT$M6lnA(L09w}?3H<1qtDKC*A98nh30^Xc z)`4g%RVVv#{Kv+}A+rR(M^W{>1Y?hus7Ot>o=kO_`FDde%X7V(qF%!Bpa-%D4V7!x zV3|Uio2W`Oy1V%J1ShU7FBVvJmlUh)Ba%KgHi3xu@80<%C@KrjPELl_sGN}%Hh85W zoUWgrMWAytrSJa!bcZ-ZC}K{-+A$st(8v_HiqqH9(J>P;UQM~-kbpfH*!Jr=b zc--xR)+m>kW(Jf_egkwNVEvfa1($F{e!&zt>dw>UPktB6We7wRu&c(0--Koqn|???4&k;7S0z)x!KtZ zlz5;PU340OJ|Hq#+l=b-?NC%1#;flnM$&^w+ah0yRRA|nUT zGMHf;>22)asn@uFIym|3Q_TcF$Td%g#i385wQLN6OCm~*R_j8%H>R;X#CUOdbWW@babuLU?@@pVGhIs)KVKPu zZup6^?9*uqr~(3*GBC*`DlOfV?^F&cH=rkeO$-X;=^BdNu2S{QbXpr*RErQup2)p1 z1s`7Z2MLteeDEM{xKA}4r>^H5w;`@7AlAr_UDbvlh@hjm_4HXQ-`>y=PvqiQoyj8v zFfLoTH>h}F*c;CbP)7v{Ge_zRT>0;>py|UzRcY+?%?)!H!NF*N0Bf6TFD?^55V)4z zd8r5}|NecR<>_p%A;)vp9WM~wv{9aqWLO6L(e&>Wsshr$gri!oKL!yVYi1!5nu461 zw;3n?H|#@+V36~kQt$5S>Qm^;$t=VlN#iywJ_qzKb32SZ{;R1vzMb;d?*MWwl?#oGh6RC-&{tw6{k^ZxmsFU1KV2J` z@{0rn1paLtsgZPw8Gu>Kpmx8=4XDtFQV7i=nd(Bvi8^<^#3p%DxVlqN$;yM80T7i6 zuv3f+*E;zt27G*cZ|mwd0;W|2JkdW%Jr5;~0w@waJw3z)95q=25Fgj~zjpH&qHrfa z1Ozr<9T!_NZUZArKrrEb4wsR~-=P=>d@8na;1}zExBxI@fQ99CG-@mt#l=R&W#JEv znL23~OPqq!hL0n=qKpFZgbN@55cbqF=`7I_)~BCLV+KOHYsi<&YHDgs3e5sMro~^H zvtu4f8fhF6csvooG_`C#Qy77sQ$GIfX%$@ z?~;J?m8}1=q5QvpggmH_|35sJ^Yu6n-#~Ie4FbR5@8G?xHkvhF`4eG)jSl6!y4Ck* zF>S8-bpOGFvn>_e@eu})!e?mRSvn*o^|qp7HL@c*>^GQQJ;RB9E?iPo4D=5ryBwPC zn%Dfi0};laA>B5>y8^?G0~4y<$;DOnj(5*1%wozRD$6dOId{9yF5rM9aE}3E0I(%u zy#JkdEWIl%VgF5d0y3i#Es<6>a>2r)$_b>4Q_sS9MRx8y=E=gXpttpwzmbDHon8v- zg0iZrB4Ppxom;5(3B6|!mSeCIxQ;2#^JP}P(8rI3WnUDy0*J?qm5`|*uby*G$t-Tm_VLw zQG`wFdh|18MJ|){!k>WD0TSN~B}Hl~KSl1u2|6&S0^k7_epW0k5*bqzHA{wm{ziE^ zgzd_~*~hKOUj4J9zkY4>@>;nsC>SL`yQs~Na>Oq?TLi;UBk~fiL+$PD#7b?iMn|mI zG8;6}H$db?w`UKiou_+;e3Bhn`oHiMlhijdqAfZng9!o5x57`bt?vQY26yLYIXMC^ z#5KPEey~H30n=+r83yM_Ep6h$8#XtjTS9^sDdqHZaRmUTNO_}%DX^4ygm;QJKFuSf zwXzWmG1)0Z0t?U~{KWh^Kx^yqw0%-iETbOyth;q&QJV|z?k&D$Vo(5|*GvC05LQ`8gkF;Y&%Y&J#$yfOgO#QUH41E8QT)%e07{|yn7P}&4_vhJV&N-!vR zyY%Qk>>h;T=4NK{vYDB-aW%xmhwy8m6CAT#?Tv&1!xH=rhg2B8p`w7uS~Jylvt~O7 z^eTjlKq-U4`-EyNbg(#Kz<5`oaR9aID3}4~cdMD=ulJ%QBohCKh^@rI{hQNG!e3It zhyq_?{WH^JuMLZ-a>`7JTw3s_JUtT>@Nsfx;zHoHJAWc|MT^p%%2m(TtijCG<^bNt!X@a+%rv+O*MLqH1O0P_^M^YI7@o}%Mcz>UR*mWQxgXnv?; ze@|8`w*d57@llL0r<1|pHbX-a4{SKV{AiPsC%lxEzBs}m#B=WZnub^Q=Ebr0XkU8J zAB(^8v_{!Li;+LP_7IyUigA(PCX#g3&>YZ)r$2kN9{?m(%XW}k|B7)B6S`px=cAc! z0PKXAacwlO^-l7aXQn=N&bdOd7=j40OJh1&@drt#2XTQ`_7hlMj)E+Q4{wB*d-&3!tiR;iZ?-SOm*##WMGqT{s?07Ow1d1kAKdhfps;H z&&6X@E2#TZw;UW^kVb>aj3_NN0{ydY@;u&( z0?Ijq?ac}U z516TJ0T+pBtFVEFb_s}_jG#y4Y#*&&<4wGI{z5YN0ucq!&nqUhz%I79_yA#EStE$@ z5UdI8LSKf0;@`xsrf9;%!h#pxa848T+Ro^Ro>N6iyF4e3L1c7~c*`PBr^pK)j0#~n z`hFM<@xOD2iFl0|bY&SHPUxoE9CcQcW*Ik0h}zhrIIqKL3?x5c;vO~<1=;+_R12|s z+`Vb^b@SV|@~>XK@`uy(DT7i^LlpXQn6xN3j&__GAk0N17sm;6L zj0u-AVJ2wi53iq-1b0%M0c7!tt;X)$IAjP=J)jfxAYX39O&7mMus*z+cXwBpQ5cNe z+$XAb6GxvOR3^lw2n>DYluY8?*Y?tuB!d$2DMmmT55ugKK6Q;WKqyZ&f z-cc>OWp)9*pO|bfd)i|ilDEMxI5C1cw_BaX+}ykZm*FPPIE9NGi5)WTsk&Z-Ipl$9 zFBX*uo8Q6^Hvn<6%De*1AuZ=1y@N;m2tGev z8YUVi#<~$bE0BjObc{WC!DwJ|(nCM*Gy!kyC~0nDIoQQzK*xPQ^siUbc_SAeHk=VW zJv$Q;&at`Z{OfPuOe!Y8^ZFq7^~)E1TAL2I?_|8P;Ik`>MGH+%w0CFMFzU$S*xevL zk>SihX#NX@!{63$0Ye2-3{Vx5HCoH*U!_25swBv%@0#d4o_TgX@LV=`jzZf+gm~%? zqXO^=dD$g_=~w7>MNoCbA|sqXEufJllu1ttF|rTnkDgJaE2k<=yGsIeWe-L~&L1qm zFk^s$lWR=@#=%jSLBbEPi0)3nZGJON)=ISEi6(VPB;8J!(ukX1Sf~kcB|Ld38pLA_ zY%qg_eaqdt1*@DJ9M> zCELI(rg#%wF}|JVhhC5|^{ukR5MDwm{Q`nl5BHvrG%BF?V3@B_6cGVhO37T$dP3R3 z>F2G^&JNykEgz|Zy2T5rA;0#|JqXvgptqXb1Xthjx_3 zRAqzZ?@=$6V_5JE!y`njys~~F0s6VL6Zq?QaIL5)#GC}ls@vimoJ?JZJW4TXn{a|k zVCwE$irnuKKHU~e91YqC_>XMBfPjhuf014k2@dVJDg7gG=3}p=q2P`0Hs+LZ*R^9u zf8aR$`Q+GhVqt_TpF*S$L>aRN}5$V99z5(aN262?vInSTV<9!m2_#7KZlkhOLYM3fI6SpO#m2kWbqj9=Kh-#VWUv<3r@B#XV|F0au?0)l)n6(qc)K}mkqgnWaZ}3BbZVg8p_BnRqfk)8 z%mI518X(x5F)pm49U%j;{|x`T`>c9L;{TM1Gc&xUn1XKN8PqHDr3};v1K_5M68l+L z&y-=FPbTj8IA;deM&QyD*#sGbfp$S=ZkUAii{6v_C3@G7m4mtwk-g+>ns$I~7c8?V z^(gKW$~n&I>BS-cW0v2nmpqjE?8S?H9alE=(ORN-RD|#a)M8>vdT5~dz%kp^<6fiQ zkWJz(-ASgR&u3muoQ*(b-id92#Qhr#OJ<$Zz6^7iTT+-j|99l1$`XiWIaroy*fbvX zW*@Z8;wA*f>n_gD2jFI4C9#BwFY^d;B`an8|HK)ZlwUOu3WHXN(`W}X0!YPuUghA> zP#=_*Jt%Q+?ojSxnlcowOMhiP!*agXdbBGRMlBQy0RzIrS94umTzOlYm0EBqQpfYi z1iagiB4iQ6S+_WXIH0mxwR-g%w0b2aC4_^8H(n_=f4kZruzgq8c_d>({Xugnk0na@ zdp_W^kn!*Q-LdUp-Jyfp;WqR0sJ6Ct4t3Lw&zauhbIm1+<@IOI*b*vwcWb9SSy+Vm zpVzS)&M9}3mp`UUbOhXBm-Uo}+Y}(?2`{jM9olgv$hSdR{ZLO5Qg2Ml-y^JHJ(&es z$E^hh6~#R^PkXadZL`EWtl{@G#*_aZfOa`LPGaze{%EV*uTd8`4NV-jZOLH68q4c& z`D2>9s(l~NF9-8)Go(>eVein@io#_sVOE{x>`|eBMmpW~&o_)pL@VhnI9Px)w%&>P z9kJ|RSUuGLP{u=w{3yh7!m~sf8M>S3g9yt7-@!O)HZznuO4+OonnRNl6K^pd`_|j* zgEo!`ixFHuyRDQcVL2WQ_?~PThm&IhLOz@!_X5W91L9(iSL8i?*}VAuAF`_CW6G21 zLT5KqtC+26R}w(Nf<#CTXe`;VLI#%)vPJie7y$qVGNXvOX^#h@9m+gTaEy-P9ReZ( zafh={KkKcLV&jGCe?PL(oQMdWIM?racBV8Ab{+xHv>{0n3nEw;FhC;Y-~vEcW@$=` zP+-tO(Z+ri+KV`?3l;;oSH4Im@Jcns{BaP*?$=vCi9AtPQxm?tS(Oh!b+<&vrfDnJ z9gkvSFtwoKUWHwT8@eKl83_o2ufYT)6~c={KTfRaP@cG)+z%%`?BOKN(8`$x3fiLZ z%YTEmR}}Rlr4020vEN#pj9vMB((Ql8D%rfvT6Wc?i_vaF*KbU}W`C*=Vm^T!h-aF- zzdMcGN1773-)?EjKH=faU`6_nw2#{QElQkvG|h-N9~OR3{_EV4TH;L!@GDsIlVwc= znX%6^12V|_$3?|skvZ?<{$!A$CZ-?SmHeF@6y6m>3`Fvm!>&$X=E6or{wS=-w~t3b zR=v2*3gf3`UZZ`oXFSeO@Fg7KzGDJT75ruABo6Y<{rj6uTu0ZP*yj)=gW5fS^NOAu zJl{%~>p_#o4UxiRg<9_Nrl0|%MgZw>&#JY;89Zgr>7~hwspU*Ibf5$m$IBx^auUT zRk(XUzT_m>VbLTpZuIV4vv=B~s|bTn3Riwm8pd1Prwp92%v8uur_rDQFZIx!uA{#E zH9KG5< zg&C`oI$iaR98+;D8NG#8mWqP1m3dReMZdo4s;W1*W_K==%@cs`ZltFR;gz)p#V@x# zqe6hf9z5_`2mmo}Xl3lTD}L)I1lT@)aX@>laHgxv+}E*bzI5SNA?+l5m4ls_y&2ig zx`4$5+-nXr2rJRA$fVGoz{O=o!w-Cw{jZCW@!95H+I4q*vy~Si*|0jq2miYfgAl-B zH2iJluRFTu0l)+qt!~({H)~hP|0ziZzdI1nz)-xMdZ zy_S^6d~)Gih|=LcvrbpdXmrjrl?iAE(9AYG^g=i(%HtCgYZ&~;*$-S^R*c}ODSHfG zfBtrr09P)&N=>0k3L?n?%tC@nb6Ch2RyRu5aWpGVik1SVeB7%>UB>M~z;+%?FT!$d zs_StSaut#_N3A(ZxaV3uzy!jJ3tZrHN3WHv60)+jPiA^+6);w+zLpi#BjNcTEPbY5 zZ+QM71d>PB!rxsr`ggr4F`C0b4N|z|!J=Vg3xavkaR)CLzSqHMVxFv*1O~#Npt>5??W>rRtLzB2~9Js@muATItQCcq2WX zTN)S)HjcGc$8ZwyRyR=!B3>;mE!i(Gy13{W(p?AsCcwx0P};>IF8`Q9k_`%O1d!mB zPq1sh+ea<8{Ofw>6D7A3N4fIXy!2XGCX3FP$9`OWXsQYG)pNAp2`h%YJ#d8!q~U*u zy3@f-zykp9c=yWkR=LaLpYv=$=+BYw`W1gRoy1%O@M_EZgrc8&bzTt}L-x?=24K0t z(4RiJX#V6wUu(Q zQeSx-8JU_djpP%sNZ;_-75Q)!?W1)G?%-ZLf1xHw+%>U1*HCi+uVlj{ z(G{2+G2Fu-mT3&l&CxN-#zg$Rm^Ytzd>O#(3P>BEqJwA@w~!m_OPI%yh`(AHDV#rt zo^WDT2P^#KGZ(OqmlQVzo`kCM9Ca62WnEAU2`9l_u$5cwnyYK|EYLVX5_MOMp?@{w zSe>SSlYvJ{Mj`&SWdS=QCVV85K{BgO#a z4MV}pJT{B23<|HI47&`^))s)eRAjvuhsqU5x}#bN?<`p;1B@4i-+Hj4{79XhzT{y& zWBxy5XILy@Y2j7H9mvBPb64}4^` zi(~-^#!RD$RK(|I8iMWl|G=3@Sjd%lb_;9EQCUK&L7Vcq#R}R4a@jy4sCL^}u_9aO z?Bqn$C};udPIGK3Aj?sRrkwbgq7yoC>eT(6J29Fdw$Dd(btB)bEx}ijbNQs zSxJ9=9u^kn3o!-{AL_6FZ~+7a*5c`x+hiA?9lZ4p{DOa=vj?;4|HVI8FY!jjlvwD3 z;EjPA@dL)=Z@^s2$9!>1LyZVAMG!WSgNF+aoo4=J!_c}X=X6mR!BbTlh{Xn2T#$+g zsO8PN9|*6n9QxN`WlaF#G2nz*c-l<`LWsAW{1i;RaAQM5(eV+r5b}|n9{Qt(d}kjNNJIkG z+6&>nJmwSCuMa9r-3J!29GScn)<5wMJc69h!(?Uyfw8G#B4=LqKpw*^`V4Bk@|@}1yJ3@ckY05lUvkT%g^Jb|eb?`vBg z@AoFwQVxA|udj~X1)OgVLSsdB;E}~76J&Ja(NEk$p*az2OOa=ja^yz?;f?4|*5jd4 z^1{3D&xKl(5^2gY+J)YXq8Y?7f9MGnR&C@VF4zVK0Bqkq6h_A_BAuI$R< zKG4eoBBGu!GBYy5PDyAt{3J%63$Kg1j(LVLk_Oq&@xV#be{jVHOBg6z5X)%Be>sG; z^As*TCPi1**O>Ky5>5vi3HD}SM}tDX`aAPyQ|LlZQW#K*?ma70l$5dx1J@^Da8bMm z2dxMv!T$E3MMQ8g1wwx?#~BGW#7VzP&>Xo?9&;rwUpbVE*x;$f^F=dF8!V^|MBNDN zG7e*XhxTr-l_1Qv#Pf9z0@BuR2E|hWAiX z(PBL&v_&UxhhI0cm8Pvxg7Yz@uVsJ58}HIpdx3`&&}>p%cx_ZDY9*?<(%wM-YBdD~ zUw6ZwZkK`vI}7#^8)YAzY0&YA_Xd`%q`#|{%RaEf$D$W!km8NquU3nxI0Cmm;&8W9Y@fW2KF8!hZzdFOL|0 zf(8U-9*EOt*xUuhB|FOdF>8-#b4D+9_K73Bd-!wy{nY9|3lZCKbV(tTre((R*a9VF zj1}l%rYApiBDzz5I5JWS`8-=hav#P^up(swSzZps5Sj0sA$mWwGs*Na%uyVpTq}TDEp!=q#E*AOS0e)kJW?-M4As#5J6y7f z?{RIun@Lm;XbZ@IqA0f=fTT8F{>YKJPCY=)zFBmMUm)&X`G#UL6i zgm0Kbqh+^0xi!@2DCpu7PUr}7CI~GFTB!+DNkoU3EnVoOSw{(jwH$sP`WX`q>DS+X z59|k!J(1_^p`}Ydr|gg4b^quct8#MTTBA%|?N`_#S(L4y;x#CCZT?2{n>=*jY2V?C zkNX@4Oe0D6v}-t+l~t`38|Gd-e@aCbD9O%#w-wtb76}Q2z)R zDiJtak3b#fLXjc&tGxu!8e2r_6rs@oqMcX)syC7dC6XK-5CJt&nC8CYJId2YmbquE z&Z~R`bW2M5d9S0iUWDXHNWtX$1DXL|n$qpM4RQnb-kY{JdYZI0HNDVv5&I^>P4-Cw zqNZ@cpFl?Xg!VSTb`wruBN-L6)8~n-!6HO?qv2KFm=O@*T3r188=3IG@I!0zwg*kn zzCfHmv4V1uLv;lJ89k@WHV|?S+V`wNpQ?J6#ntKe!9DsNpN@%s^EwIygls67Sc$(I z^9PHY@F6}A{s9ZQ62}kRoimBH8p}puV@KTHA>-{YcLdJFAT zRcm(Rs5CojyE9X%a4_Q}XBN5(1_0N9XlNI-pQ6I-FQ1G~f1TU|LXYMCj{uBt2b~(N zZe$SVk%UBUTd}u-3!cm2j(jcaZW&!B+zt(t3h=Gc&p&XG*~ti$R#xe2qRaE2yv#lX zb3eTOIT?}&uJA!8+{%ltaOkCq$o^Q+gz_CdbqgY z{gIDI%4UFIMDj?ltduSI(XJ8W-NboBW@fOq@2~vuxXg<^%7mNuDZ?t3eoNH4d*#o2 zdwO_`3t{ObB(kwa?}kLg)m!vzpnqXli8W@B)Zfv^Wgx`?J}+o4xf6WHkeIF*Y|&FL7d^h_9eI(dFVuQ_3c+%c0_$ zy?!nHRE0@wby%`yOG!|ZHqovV@i&DiBD>X3o_s?vQ{*PG5Alred!Jph`o%=0B1)ex z;J1Ni35;MKL@Nc5wnHXS1#Csgp!NpuZ!=na!d8k$j3EnJeUo_2jS|PcW8g>JlAuRS zXmcU7C_jUJ72RkkRmKH2D66Tlz$GE+<&pSBD?cuS9J8BMlQJMp^~_?DvY%|~>BzSe zA@K+fTdF0h@kO}tF$mkR=|QZks0JU9w+}Ci6>9AGe3}_A#6j92PYtJ05tRuLy}Q`1 zK~4S_iC;z6OngA`&kp)K6j-N5=j986n2CGi8g0uoN3_A7P-rq3$!U;bJ(+X^S7Zn}HN!0jPzGp6I019-WtUCwFJibY{OKuN8QvkpsNNT0t_z$x0@D9ld|}^vdv~ITtI61w zG>StW5A9XJ4o{2^FuFV4qL`qg4Veze9E^zCdvk2b7`dUaddC^T_0k(_DlukVMG?C) zP;1s-e#=?)MCMs@$Z)`uq5%>dh%zN=S2qg7s#c)k2=y;dP9(i`2`RY+3JDU#cC7ti zL(;ni=<6t*=7oNNzeb6G+(I5dzKdV1nuR{3t|2Y8_9u)<(UM}fiRo?IV{Xf9DNx!R zcGFJ#j9bWoByk)_9_nO+40qP?hal2V6A|!FtP?>x+&lh3dOi)?Y}^?JP?Pex=!UKv zQ(Pdc-ie9M;zIXIC7c^DT6izMSN-H`a2kQw-2mI57jOYsWM`zg(v}$5G_6s*%>hfh zg&j>#unK|qWegy7!Q~&1gtS?3REW*x;+WUUF|3Zj!93GP-b_7&!8Di}6iwD0Mps(B zVlyVw8~=Jj#(TwN8_Q;eHvy{!0&b+HrWS5q&(Uu#qWer~vli2Svv!O97wN^EeZJls zVb|Pqd#}BJJU^}Ebq0kGsjN)F>{}F=BClUJt(&}I`AhutfZ7$xO1I?Q2(2Txes{X5 zH`sJk7gw;EqdT~3)Wo;AKi3)hc_k2dK?NDd-DNL4NE;FwT0R=-h-DL)jIZ?`<~ES{ zw!4z(7W?}9;n`Pf;<5;kF>;^f-Rk|rV|}U#o>m5=XoS!lTlfM{wi>2OsOW0Nt}$7) zTkTO9tc!>j@tqw<&i$A|#QN6Xzq-I+BZW8~qQG-hi?B+L%PR*gO&r3AZ&Uv+1y${{ zXNfAxb*~OmV*{4U6ep5(4w$69#e#PzuLaj9g8zBONK6)1Jr?Wt$ic%P=Xq=Nx8)L^Ks#beV~{O^z^(%kxTzT9e#CJ?Cnjl z2+IUUj(Y?I(kfs2`?9i6m-{ISN4!8UAV#?a9vpi`((XlCp72vx_Ok{WCX^IbM`FE@ z2FT0^Sdj_j*r354=^?DqZfO(GBbnI&4KFpT)+C$!H$i8v^6!agE$d~i;GUG zQ=qIPxTYVv?FRFqr!10iUr|Ys07Fa^WGbjpdeQAdsco3~om^XVNz|$5+O)rfjJcEV z1~Xxzf@uL$Xbm>xg5!(jz3Z^A^ewLO^yB0BWH8!WtEHs{r^+k4=R4Zl?XxFm;b1^S zL}M;n;uV&jeHYlXhmcvxMs70TMABc}$_$HLkm#q8T2vu z9gf4!d#>tuAUKCV!}~K|ck{-=0=O3Ki1Sf-dHeODqMQlmlFIYpMkx|TWr6?^cXU$h z0SrSgSX#bEnrjEaT+glzxjk#|Dr%}$CHL@%h!+?~h~wcNfTGVZK($L8vYrftt4oMp zuBy;Y2BNag!2@hQ_}LCz2m>!pcv~1v@+L|R-`b#l^e7chA1AQ-=V#ZvZj3ji8AdZi zvC4a*%ol13#U6$%lCzFl1^;YLJy(?Kh%pK1DOc<^gM~q8Cy!*k9v|LyEav%haYCL4 zGxrS;+G`?Ga09Bqv|ZG;Ljoh%V9QZ&KzwAmI=UWaqZ0Qjp*rXt$$~X9`~nRmQ!?Y_ zv`dWes0n}d@@J+^Mhno9ieh&y-70F-{4}67WJ)f=;j7>i?i7Tk5@SUZlas_Bln91! zHqJ#v6~&NU5Fu$Nlb_T2&n_HrE7mn2IcAcX2iN6q}Ad+$ zp7Zz`v6ny$793~Me(n4Sj8&9U^mtqZge(iA$4N=NxH?)KBA{{ecEOYJb7g_k$b(al zVCM8RNbB6WYANTRA1ZS4@-kJ0Mk>BeBXU8>d+Owk+z(jp(l1pB_$ZVXassJ)qRPX zg+=_jT-o$2V#rX~ZtiTGr%U6Qyzb%S^TW<6_Co%1hk$a>l;LuuII&B7lYD|Y)4r5J zQBiT)FH!qFDw_u0_}V5Gt;` z-xw!7VZ5boL|k%io7>rCA5g>E&!CCs*57$)pYRyy8#n$5d+?yL|1h|aK0jvp&|Las zJ)N7NtrIi@dB3r|2^)-bgPgJJOrqxF$NrcV%x$g9)N`E?K z02c47Qq13s`oAT$T5EBF3htBZBXeeqD1ybWdHOkmVTUD-Pml)CyZBarf&t$hSTmye zU6WTC19+UIbZ|bK=h{($H6X@-{Ja%=>rU(6{o|>QOF_C6vW3c4s&7;RAB5n^#O-G| zcLdB0$cm2gGhLJp8buLzOj9$Y(uMuLd>rXkpvicK2?T&D&Z1r;*Z@H8Qix1q`|YNPrnta!>4Wo z{HP16WJhLMZsxs`FE}UHZQGt*htuC}5>fFSg|i+$8he%KjKJU|^F4OI4GT>f3>iKcHeG#qzw%Rgxet{IhH&f^psF*z$jv=Z(FKm5 zeXzM(f!YL)P@j`O(DFvq=pi=m6B9E9ki=eL{?p6O%IRy%xtP`z&l&UtU1O6d!>^d$ zl;>7fSJ!#?9srm20Tw88LOKpMZ>@-qj(+OV6lZLx4FO+N|G#C00N!vM;Woz*30(xSjVB<)d);Wo!_i||13qE1?q_pSu zz`Awoh~E}*E2tC>Fk*cnLw%isuQK?=E>^I*F#mu;&xYXDgjPs-f&!+fY0J8sAmH^b zs}m2~OhfQW|=)0-W` zUoS#4jDQ1>|LH#W>9lzHI0m>~Uz|+cah!|61O^A}6=M7MR|In(8{Uv@eI#05)&O=& zW7e5GA$0h&@Hc4`_$1jZl3r6=+t(q_7djO7z_~O?ArtQ1|HN`MdB=+R5fB~HE}|*Q z2~2q5rx6RT4uz}&CTjg!Yw-s(+w)>{0CeWs)1NHbeAU1FCAVt9p0&17GcNSH6 z{R>NVGz4#2TO(S_y-|MWJx2PhEsGOcn8F@Ney-Lt9XXf2gWD zik<_NCIdVBI^_G9$@?|^gl`9M7ix*WmRE?-of}T#aFH=?W>a>iF`@E7a{gm@E0l+r zk`l2O^0MFU+iUN_v=8Rtsj9NL1q%ko4=0Y&c+#YK?gmOO*~_dzBOXPMISkRI;hu0a ztqU^T6(bS$2RQp`d~0SotoZEI8hOt~;ZZ~uj}$bmpLc-H697Hs>)70F=f6)K%3~Tz z*N?}np$PG)V$kOKA=qJoFtNY_5&96Q_QWrI)tWVMUwYf7c*fMsj2bm?WOTHHWz68Y zt=Rnw%cOt2boiO3H-5;(9244y&# zVg|Td2}artWGYw&vw+}A_8>p=Sa6P+e2sw*D$|>&!>K4J+;3qJfy_iCXcz60Kc{6H zKZuAZ$BZ3)4P@$sI^hi{n@dWH9gPP(^2vN%6lFr*kd85Z=HlZ97S$oKU_8GovEgg} z*?%5HKRD#w3Ju+Y7Vg!tWcGBw)3Fm14y2w$dxbi2z-Vn{MFqF4EGKZo)`K613z^%2 z%4KHwavsYSur9*et5`}|Oujl+?}uoAHMkLC{b%r?EP~ZvTj5m^Kx}ZNYx+jJK&c1U z0F&Py#IDY`V}CM~G2;cdmo$jS&U_EDqMrub^9E*J+z~ZkUR;O#i;UGVLiENS2}<+I z4@Yjs+Z68wqplqDH82>zfooC@XSN?c;x;1@0K5r&hRTR6JBE%2h4CY>xj|r{%(Z4i zrLyoSDk@D&Z)k1lbpSar3!NB>?|k4Ji-whln>$W3Ec_lmKN<@v#Iwi)w(KwocxR~N zxmF-n$|~M92C$8z$gT!|?v5}Cqvn*~<1UKb3@{LV+u6D8bQ)P&P6YY87}p|I6Hr39 z6E&HQiq`4JcOHj@Y40>dErZL+p3#|iX}NQ_*yxbxgPMMvc+a3YgnnoZ*WxH8IS}tz!QC{03Oom4 z5eC0P@Csnr-!#2!r=?T6>S?k^a&od6p*cKNqrPX+?zpw9~*!0qN+f+fgQ{XC1{cPWY zF&0+Fb5e+2Hn{4%pkxL#CR>1QU3|ButRh_)DXE{YtzIHc+3@*`7X<~=0AwW`LpCJm zG#{Jy?!zGhwY$G7EI@&WOk~j!Q=VQHwHdJi!J-WlmmtZ5ptaD^)Au#1sjA+@2|}Gn zA9CW|Qf<8g&t?=nAs~ALPohaZ_k|5k_QVn1^TiW)8w%N}A{Bm{7R@9DH`(ZQpPgQM zj9J4<8N?LC>*lgd+P{v^n#CYf!Sk?wKthEqInAMoee-52gq3iY3Ls{gVxQ(PSOZho z*5}}ydg$xV7m%zmSbMWO)}0!o&|_O$P1wzkgwc4yUX7qSa?O_|!MM_W?I^G{3#fPA zOdddSa|Ax@`OHw8|7QRN5it$IB?5q7VUv?5d|sRlr6N~o6$GuHw*MlIIKTleN{qh1 z59qY)zp%X{{?H;VYIg;3tcb8MZ!tdvIDk7U{qN*upwER

tkFVtks}9aFUV0Gu{@KTSD`iV29p)>oqQ z^YemU>`zexomtwY*I$;z9q8{*p+Mq$+fAyPwH))^O;~rJWZx8@Xk*B-N7ozFJ*9mo zvXsr!kK5ae5tBZ&ek62vPPhK}GIDAcBjL9rlOurLii%gnyk0s?SJCe_ma9Oyiqrk%g$)$8qR zL83Y!m-b;K@w7l-K!6!;S5 zdPThE13R8|AV)0XgUlAh$$tM-Tl)@x5LxjaF5>}32c-;HERlc^J7AIdlS!w;y<&%f ziufTSw@B^Rn|?LU9omF~-4lO7QSn0j70mqs%dOAr>rk z>$QO%B8A3Hyfgui?Aq_1)mg@+cTq;BK8i@30t0qE$r7MX2Bi$S|KZh`9;gD<-(q_! z3VRJ*U4I)10mY1j{ zBc)x$Q5)Pb+$H~E1ykN$8Fa8+{=Ay%hgJ_v3s_fA(vV&E82wsFu%>+kxF@S&-pQD$@R&mVsU9@LzviTeN~AR%J0+#3)DTGAq_ zr>TjLTbfo1Mf(Ur!#Kd%rlR0gDsWt3409&I7#_$BU_3g_)IfZ?4#(}5j}Oh*YwZ;y zF-gfE+sB;v?)+|pen=5yIV_CW032m6AGgAG6rt$Q$pl>Rc-9vAJ9a=j=cby}skDz& zAX>zRC?^&G!ohXI^nHiam~x#C4k*g}&EUUaJY9w_#dcecEl)?C)b$PbAc{=@ zZbw^euXgg}QP5Y0-6to9n}rquvLGy8hej_upM0t2^x*cw6QN=5Q6O|%|47#2q&B)1 z+M|KHE+6;p<=I^d+lNJbB@ED`^{O56fS&GAOiYkLd}Tsn;woTpZ_tsmFID|Yw2ngI zOPJhTFH2x>v>6F`OKoRYWTo2B<Qn8-5Xy;{=KK@Zfl7!a3K?T845Z#nFn>Qy$`VWsLBk$h-MVf1p?qsqr^>?}Vqh zU+om{yGSC81Kf;s-q5D_oqy}_u&cc25vg6i4Dz|>0mxnL_Si~+AIOEiMv0B+!8kpp zHgA|Y>C;q?TI<#JK(y4UT2jFI@fRGQNEhuraG>P_FSoA*FA8T=$7F6Pe%@cry1_e^ z0hUkK!FeL06jXgcu~-Wa`Udm$0DVPcrwbuw-N?LEnL0fE2j#dJxFFuZ!4O>dxJBMy z82#{Sls07Z!y^{knv>)Y5rdUrB-pvRYgj_m-d2GZHbX%!4^vVo+^CSWVww_mFXGMl z?;vFwaX18e$;}cNWVlWK#7=^1lldB$4&}hkro`LU{QLFcDYhSaEHO-!fsAmXRIC(c z 1$ are determined in a recursive manner. + +\subsubsection{Measuring forecast error} +\textbf{When on absolute scale (no log-transformation)}: +$$MAE = \frac{1}{h} \sum_{t=n+1}^{n+h}|x_t - \hat{x_t}| = mean(|e_t|)$$ +$$RMSE = \sqrt{\frac{1}{h} \sum_{t=n+1}^{n+h} (x_t - \hat{x_t})^2} = \sqrt{mean(e_t^2)}$$ +in R: +\begin{lstlisting}[language=R] +> mae <- mean(abs(btest-pred$pred)); mae +[1] 0.07202408 +\end{lstlisting} +\begin{lstlisting}[language=R] +> rmse <- sqrt(mean((btest-pred$pred)^2)); rmse +[1] 0.1044069 +\end{lstlisting} +or using (look for the «Test set» values) +\begin{lstlisting}[language=R] +> round(accuracy(forecast(fit, h=14), btest),3) + ME RMSE MAE MPE MAPE MASE ACF1 +Training 0.004 0.096 0.062 0.012 0.168 0.939 -0.068 +Test set 0.049 0.104 0.072 0.132 0.195 1.092 0.337 +\end{lstlisting} + +\textbf{When on log-scale}: +$$MAPE = \frac{100}{h}\sum_{t=n+1}^{n+h} \bigg|\frac{x_t - \hat{x_t}}{x_t} \bigg|$$ + +\subsubsection{Going back to the original scale} +\begin{itemize} + \item If a time series gets log-transformed, we will study its character and its dependencies on the transformed scale. This is also where we will fit time series models. + \item If forecasts are produced, one is most often interested in the value on the original scale. Now, caution is needed: \\ $\exp(\hat{x}_t)$ yields a biased forecast, the median of the forecast distribution. This is the value that 50\% of the realizations will lie above, and 50\% will be below. For an unbiased forecast, i.e. obtaining the mean, we need: +\end{itemize} +$$\exp(\hat{x}_t)\bigg(1 + \frac{\hat{\sigma}_h^2}{2} \bigg)$$ +where $\hat{\sigma}_k^2$ is equal to the k-step forecast variance. + +\subsubsection{Remarks} +\begin{itemize} + \item AR($p$) processes have a Markov property. Given the model parameters, we only need to know the last $p$ observations in the series to compute the forecast and prognosis interval. + \item The prognosis intervals are only valid on a pointwise basis, and they generally only cover the uncertainty coming from innovation, but not from other sources. Hence, they are generally too small. + \item Retaining the final part of the series, and predicting it with several competing models may give hints which one yields the best forecasts. This can be an alternative approach for choosing the model order $p$. +\end{itemize} + +\subsection{Forecasting MA(q) and ARMA(p,q)} +\begin{itemize} + \item Point and interval forecasts will again, as for AR($p$), be derived from the theory of conditional mean and variance. + \item The derivation is more complicated, as it involves the latent innovations terms $e_n, e_{n-1},e_{n-2} ,...$ or alternatively not observed time series instances $x_{-\infty},...,x_{-1},x_0$. + \item Under invertibility of the MA($q$)-part, the forecasting problem can be approximately but reasonably solved by choosing starting values $x_{-\infty}=...=x_{-1}=x_0 = 0$. +\end{itemize} + +\subsubsection{MA(1) example} +\begin{itemize} + \item We have seen that for all non-shifted MA($1$)-processes, the $k$-step forecast for all $k>1$ is trivial and equal to $0$. + \item In case of $k=1$, we obtain for the MA($1$)-forecast: \\ + \begin{center} + $\hat{X}_{n+1;1:n} = \beta_1 E[E_n | X_1,\dots,X_n]$ + \end{center} + This conditional expectation is (too) difficult to compute, but we can get out by conditioning on the infinite past: + \begin{center}$e_n := E[E_n | X_{-\infty},\dots,X_n]$\end{center} + \item We then express the MA($1$) as an AR($\infty$) and obtain: + \begin{center} + $\hat{X}_{n+1;1:n} = \sum_{j=0}^{n-1} \hat{\beta_1}(-\hat{\beta_1})^j x_{n-j} = \sum_{j=0}^{n-1} \hat{\Psi}_j^{(1)} x_{n-j}$ + \end{center} +\end{itemize} + +\subsubsection{General MA(q) forecasting} +\begin{itemize} + \item With MA($q$) models, all forecasts for horizons $k>q$ will be trivial and equal to zero. This is not the case for $k \leq q$. + \item We encounter the same difficulties as with MA($1$) processes. By conditioning on the infinite past, rewriting the MA($q$) as an AR($\infty$) and the choice of initial zero values for times $t \geq 0$, the forecasts can be computed. + \item We do without giving precise details about the involved formulae here, but refer to the general results for ARMA($p,q$), from where the solution for pure MA($q$) can be obtained. + \item In R, functions \verb|predict()| and \verb|forecast()| implement all this! +\end{itemize} + +\subsection{Forecasting with trend and seasonality} +Time series with a trend and/or seasonal effect can either be predicted after decomposing or with exponential smoothing. It is also very easy and quick to predict from a SARIMA model. +\begin{itemize} + \item The ARIMA/SARIMA model is fitted in R as usual. Then, we can simply employ the \verb|predict()| command and obtain the forecast plus a prediction interval. + \item Technically, the forecast comes from the stationary ARMA model that is obtained after differencing the series. + \item Finally, these forecasts need to be integrated again. This procedure has a bit the touch of a black box approach. +\end{itemize} + +\subsubsection{ARIMA-models} +We assume that $X_t$ is an ARIMA($p,1,q$) series, so after lag $1$ differencing, we have $Y_t = X_t - X_{t-1}$ which is an ARMA($p,q$). +\begin{itemize} + \item Anchor: $\hat{X}_{n+1;1:n} = \hat{Y}_{1+n;1:n} + x_n$ +\end{itemize} \section{General concepts} \subsection{AIC}