ࡱ> @=h>2I]Tto;%z\$\2o26xڭ lVǟsR,Cxa*n|Z[-vldDO*Ә5Ekd~`D730U1!B4_Ć Is{mӷ9=y{93YNCM";9wU8=C_˔]e#ctgL_&9wCmVnӈ9I}KenLJԸrIgދ.i4W=<ķ?~|'6u+]M]~ƽN(OV$?P㪵ꁇ>6nܚ-21#7#-Ӵf͚8b?>B~6J3$>8 oE|{C|7J|Rdל?ZU_zf鿒F`$kw,h\[ehsacJ *,?lb2.+L\c!H=I,Aw<#rcRGxUFA<9z=4RkuI?pC=,u]wNA]=~lck: ?ʷg8L. ~ԍ(oy4~R~L>/6?GG͍~>?̍P&krx պ='g_oKɏ 9g,썍q21NʷǨ+!!bԗc_/H!-jy@L?4: ()ּcX^Rc5X%Ǫ Ī*)VCX͗Jfpu:I- q y\>>=|#Le]΃uf5{ )1i6 K _rá[o%Jxqm,6:hYd84c _e9?K.NKý%& AeCof⛉w[AAocphmZ ߗ  {- $1in_7Q(IwR?E|_I4|74  dד^zWhB55uuR7Tm_ަe?4ߥ.}phOcPP}4~h'?!~=J _?O3Nk?/5BGίI~vi ?gO$>dg?9.{~S)  ^/4.b?uIN1i.;? ȟ 7o"D| Cw幆C5?h_׸0& K ~^+ȟJg%ꯦ눯#~?\2&b746ĤR}O ~>@OOv"N>H0] ܯ KI1i]LFRá{JL9yyg={e24՘4><ĤRphn27]=H$?D0?$!#. )Oi5Ø41 %k1i;4I$)v=ގhJ_Ieo'7(yIOd$& 0QveϠ1k'r_FQQ ;Nȶ`ibƚ@= :=>0bO<z$\7 xݛlUϽl,>E"  "PBc{miK?^;8`nRk3H6Fmfh܌d1,.Y da!sİH6Mw{8볯c{~iuRԍoJun KK+(+5D)ҋjl8\oWt߷tpDP+jVSUک#*jTڮ]*Ԏ욤GzT5K5ZƟtYN]o5#v5:_n֏8f/~X4ݭ3DzLx7nW2?i gƫ窻./)C|wǫ3_r,Quy|ڻTuw^Ǚ竻u뛟kO9џg˭L;RwnR#tNݦX*_dz/h'oF =p3_ ~)>/AO?,Ծ9r?@.v\uZo:?j[Ϝsq{?:jLK>JBac 11˼ks-<(ӷCuT^>^{]c ȓQo?#f (rWPJ W `QLpc4,A%N L'uȓlԜ9#ƊG/cˑǞ!1oyR<.ݢ7y۰o9ym^՜;Ug5W1slu/ü?lUyWQ7?><]n'/r婿۫ G݃?__*W ~?/O^;rԟuk|gz-M|g+_N?,_~/_  'Ÿܫh;_WU/Z;>ſK/-yE?oşg[7#|/G娿R/E_됿FYhv2L#--_^d?ϟoA>-g3݌|cC-D߆?E/ZzV7|EDaۃz_/^O1?0. \zttt+t+t \:.ȥqp(t\ \ \:.K{{kk++qpt \:.NK'Qpfpfpzpzprpt t t%t%tt[t\: .NKcғ'K>.] .] .].]...NK'Eqp8t3t3jpeeKC1.NȥǁK .\\\z pVpt \:U NKIp&p&pt\z pnpnp --)pTt \:.NK7K7ȥGKnn...@.NK'ҍҍr#Knn..\\@.NKG#K?8~.}\t;t;dpɷKwKwKwKwȥ#pjpjpt\\\\3ƥIpt\\\:.KG'KO\\\:.)KOK?.NKKȥSҩOߊG'p[-ͤ@=< ;i wU^}8={TJ\7 xڽ pTh-p! 1"bI%mx B5B(MI)QZN_N>|v|H(L((ժs/03{Ϲ3Ɯ噎#ƬU L̘kK'1WfϔضL>n?c'Na.n姙k̵Զ,Sdn1 ͦL2Kr~zZiS}k[ɾҾϵKYmNo_e2hR'5W}ϼj2<]F;I=ܴ2o\MPfShWّΌev-82AO I0C̻oֺ26ҝٴMk:i{JcG{Ŗv)-Y?YΛk+2 _c^>~j'5vXiO p]mU:c\n?4:̋"Aּ|0s?f1% |&'O >XmYW3ݶm}N/Ba;=h,g=JTCN]LX'/W31dͻ#1±D±~k0|;t۱Ij5hb Fw/`/'p=?LP].b/~.{kpm^\;;I v όK\dx+p=$XOci#44>,_zx4Hď7=rf둋FN]HWEXa]Ө  y>"ǭۍ]dBoNcc^^p(3% S^-=IYaCR;#~IGtއ-?}ONx6}VG|NǢ}Ik1.>:}>.MGHOuu{> ߻{9*ܟm1 d[bCY1 13f/\{tݓ| ++`|K0eXeXPPll+羓KƶQK]C(.YU\7}Fq›~MX?wS\m?a؎oGؾ;}a|9:[qw6Pqa[⒅?]K+.Q\Kq^Ÿwb%k |'ʝG߻KmKw=^›m;d{mkS\pw]yXq? ~(QOvPq# wN>Ú?Eq£Q4]⒅wUq{x,% Ћ~I.\ ſ53$ `_dzd99\Sqݱ,% wIeG,ܝk(.Y^V|).YDQ\I6?d6?d.%  Y /8¿e,|).YVV`G0c6R\ټBq߳V% [,#LAq](.Y6/U\p%g xPxH}dc|%%?o*.Y7"|dᅑ9~,8+YuX>S\#;eLT\~4Ϛ&(>K>汊Kc6_US=J7CQ<#K䖣Mp'[vl`n_WN9-|8r8r8p3)OS.S.S.S.S <N9N9Ny0888/SΆSΆSΆSΆSΆSΆS <Ny 8pᔳ3333)SN/r_8888888j8j8j8j8QpʣG)S<S <Nyrrr8pppʙpʙpYpYpYpYpʙpʙp})_|r8pʟ){pmpmp)))^8pʭpʭpmpmppp'O)S>   Ny8pʇ)S>| N 8p)l;OvQ&)n&)gFSό$k|N^8po) "p)? 3pʏ)?Ny'Cp)S#pʻw)S Ny7pʏ)? V8pʛ7)oS\\\\Ny=r=r=r#r#8-p[)7)7)7)7)oS\\Ny-r-r-r-r-:8upupupw) Nyr=r=f8ppp7)|pʷ)S^Ny% 8p+W)))/S^\\Ny89pʳg)φS <NyB8pʋ)W)W)/S^NyiR8pUpUpʕpʕp)))))))OS\ \ |=pᔧ)))OS<Ny*r r d8pppppppEpEp')S<Ny4rrp8pʹpʹp)S <4E|/ 1$@= oe16V{Yg@Ac? x՛ lV} N@-C-%Y$֠/uHc`0 )R5d&Q:; +KLⲸlq.c1ѱ?޶oYOzy9>w^#ܪȅޔs5zhsW:W\Y5޹ȍ-wnds{X|=#;+ut n-sJW%MsMplxw[(e7W{= ;չL/.9nԛ+[ŮTJԩ[^ծV_'ۥ@"=Z&ǔo%njvML4I%{ȾYlwwG߮NJh v]<HӧO}z.i#Q_~l4kĨm<7d^>\|~;HߟRw[M^;8}=%~V}oZ6Yλ7sEݷ˛fHm:=+5Cɋ2o1v.=q'V~ %R&cJef.罴֓~>%Hts*_ w ,_~?sg߅z/~-Z7ÿ~ s[Ya\5ZO~\Jnfz;7߷'ngvw|K>k }rg&WHic/CE& C.?K*{oqu}z:.93Y根s0]>~KespL9z8d;9:rj~M}R+ק9 Nv2]*w96|lO|~qYa9^{:ɘ Sg񳘏ԹhfQܐe>|8X; s|cS/``_>Ksht'%G_һS3ĘS(Kbܜ%FYN1FucP$KA9ĸ^t#Jb#)FY1$cG\.1Fv$,1^!ưUIYb)ưnc4'1vgќSXx$KU9mĸ;e9|''1g19χ%1d1,ϋ(;wϭdcKmoȷ7Xꢏ}uD-ܞ lyw3>vqZdl{Fg~}2L02yHCkvV\/ ~%گDOAu/WޫL5jkA/՟9Rg|dV~BƷ[PiQ&f8+qdUϹUZ}U,qV1)-Wx0j{a|k-.OK{4 w/ίiZHh|? Wϣc|$FV>c|QʵW _e_L66o\ ɫů}_L8W.fkihgU_q;y6'd}[_u&~rDUύZ}?5j Q/U?]eίj}fO^-̏2h :p پ߈6!}ihCl`D%VlC"E6It_u.WsƇo_udY萿UE)V藍://UEtU2+Dޮ$:/QJC[,8_#zU%YIC%i萿s?!+ foszKZE_C%;Z?DqN__u_D0j߈2 ѿ0jьufUW(y!: G(zU(!C'+[:/wWm}úA*}!A DoK WT4t_R;_:/KU5:rs-7~2#S;u:/:/> ?G?IWUE/KC`|_-bp_,?Qo׀Blh7ꐿb擆6Iu4t_DϰKC |nYuZ_TS[R13VOi.;l\tp.ًp 3ri̵-v r`ګ~L{-Z0`iӮӮӞ= Lv0iOӞ= L L L{"D0`S{ȴiOӾL0`ǀiӮӮӮӮ=LLL&0i/ L{;v0` kkiO =L{ 0r0r0 0 0 0 0`ǂi..Ӟ= L{80`#KKiӾL0^`ڽ-Tm`imj˴S`)0|0|0....ݿL)0L 0`'Oi!> }LLLLLLLv0`ӾLLL{8p0/....ӾL"0`}iL#0`'innLc0v0v0`i>}L 07i~LU0W_~L0_ L{'N0m`[[[[i~L90V0V0_nnL{0`[++[[io~ L)000`ڛ[[i~L0V0V06060`;7ioӾL0-`[7ionn~L!0`[ׂi^LLL{'N0F0F0`ګ7io^ L~0ׁi^L{=z0M`ڛi/!ӮӮnnӮӮӞ =LLL{.\0^LLL{&L0Ӟ=L{>|0`ڳӾL{0`3iOӾLF0 `++iOӮӮ.ϑi74@= ]XX%DxGl?2CQTd8 x͛lǟ&EY]E&¯Y8Ā a0:C@@@ew~ڲRnFB$AtӥL<{q;|r<{]u).R[]7JxTSfhcnG]@{`aPB_yΜWG^'Ӊ~lA[r?Mӏg·Rى~Gr?Z%M?9s6E?u;я^z{u~>1s?Yϫ17Ѳp?$ؾc~" >/9:9Ƨ14t"G s3Ijϑ"5Y1z!m}fnqNecӬw33;w9"w(Nc{9˝SXawXYzOcVmrYQ&?*9a99I_9FXo;u\?C^douWi6A9f}qW64MJ'HQ~<w͂w5 V Kڳj3 d88?c;ZآOdO#`b߿~nǟN?aPğs?tE&kqlѿ/@z[6[ `W Y5_g}PYآo2ˊ/Lgbpc+آ{?-]lwqŶ1c[Yآ;ganM-k-]l͹tE;#!W#hb^*#nb\i.K?_,]l뱿뱿Ͽ3.~Q\xGMЛ 3ߌvD6l??"Y?yS x K%-oE[/CSٰmq\Sa~6v%آwK[[lJ?b`"{AJآSlvm3׭`[-!Wr?J>}[آ ڕb>ĕ|[آH3G}!1ڟO>OAOBt:?3? ,?Cs%[ȿb.] /@ =վ-=}+آ?#ZԿK⿂ycس}͠)пvgw /Ys+..Gѿ/?0Ώ0Ώ(EbC* 9\b~1XX7Qߛo@߆6FoDr חXX&oBSkՌ gpcO~`Σ3AK/?[>-XV׊.#e+W;G~!YdEn7?cآ}Oo_|+ԷqLv <n2FK8ǒQmHF :_3iMK=3``u`u`M`M`U`U`{MWMWMG#`>itl: 6M6MWMWMC`5`5`it l:6lztl:6MG#`1شl 6mu_6-n6E%}agπM``/MMM`tl: 6M{=```ӵ`ӵ`q8tl:6M`1Xٴl6lzI`~i?tl:6MMM7M7M7M7M`>itttl:6M`Q(tl:6M`a0A`^i/n`A t l:6]6]6>`^i/شl 6lz?t l:6] 6] 6]6]6]6]6}lt=t=f``d!ttشl6] 6] 6] 6] 6M`Ӎ`Ӎ`GM{`A t t tl: 6] 6] 6Mc`ӥ`ӥ`ӇMׂMׂMc`>it t `ӏMWMWM{`5`5`tt>}`ӕ`ӕ`tl:6M`iZ`!شl6]6]6]6]6] 6] 6M`>ik`ӯM6tttl: 6]6]6MMMG#`tttl:6]6]6MWMWM6:t=t=tl: 6Md5`5`M}`itl:6M`tt~`t l:6] 6] 6M`q8tl:6ϐMJ͇@=f~iljM"5fTd84x}hUsKZF  ؛cn)K 7v_c)!H(|A(@ȿ P;=a {^)(fzҟd;U('wEEJ5_\G//CA gSo&snׇ{\J׻V[/qAk&;Zdq͸V;tB:tzS?O+*Aų2q+t|qVrn=lq8#HcQc+21e*Cy9btrR k93c]d)5uw :.%X'=bw}:G%WE=cEt1Xgjѱ:jHU/ZxdyNjT\#Z{tYFU}~G=Uڼ}Yo #NK\"`zi^gԗ!c l lǬJrUI')9wnV hR|3֋W5-~/Eo/Z96-!N/Z]? Zmkǫ _w5_K@oh/Zr1Woh߰W]Wse_Xm-a80⟄9gſ/Z_9E _y|S7|/bE_zE⯱+\ۚ&Cڙy'ݻ]ٲS* UyX*`UUUKUKUgͪfQr(YU *VuUQ`UXVŀUq`UXVŀU`UX5KV?dǚUϪ=rd2<_DcԽ28.c&I& &m&m&I7I7IWIWI `R?0&ŁI `R&%I `R&EIiʨIS9uJ3IUcT&)oCwGi/xvWfj^~_~_ZW9^ŁW~~دu6G~\݇a9x_;k'}--mmŀ1_r1C_ +=lWd'W)5ݗ=eұ/? 7l [끭}־^h hN`k%'[ S"cf2( / 0LDArial 00TTsܖ 0ܖ B .  @n?" dd@  @@`` P)   qu I"#$%&o"$>2I]Tto;p"$:=>0bO p"$;i wU^}D `"$oe16V{Yg !"$]XX%DxG /"$~iljM"n< 0AA@3ʚ;ʚ;g4OdOd 0ppp@ <4dddd w 0Ts 80___PPT10 pp? ,O  =,1Forecasting with Intervention: Tourism in Croatia< !Ante Rozga, Toni Marasovi, Josip Arneri University of Split, Croatia"GZ =F1. IntroductionTourism is among the most vulnerable business activities. It could be affected by political crisis, outbreak of the desease, economic crisis and war activities. In Croatia, the war for independence in 1991 affected tourism seriously. The number of foreign tourist has fallen more than 85% in 1992, compared with 1989. But, there were another interventions: the military action  Storm in August 1995 for deliberation of occupied Croatian teritories and NATO strike in 1999 against Serbia, connected with Kosovo crisis. Although NATO action was not conducted on Croatian teritory, the action had impact on Croatian tourism. rrO            2. Methods"LWe have used several satistical methods to analyze seasonal and other variations in monthly time series. Some of them are empirically based while the others were models based methods. We compared their performance to see the difference. We concentrated mostly on three of them: 2.1. X-12-ARIMA, developed by the Census Bureau, U.S.A. It is empirically based method ( ad-hoc method ), still dominant method for seasonal adjustment throughout the world. 2.2. TRAMO/SEATS, developed in Banco de Espaa, Madrid, by Gomes and Maravall. This method is popular in EU. 2.3. Structural Time Series Model developed by Harvey and others, computer programe by Timberlake Consultancy Inc.l                    3. Results"eWe have analyzed nights spent by tourists from July 1993 until April 2007. Figure 1. Nights in 000 ff&  BFigure 2. Seasonal factors extracted by X-12-ARIMA and TRAMO/SEATSCC|    5Figure 3. Final trend with X-12-ARIMA and TRAMO/SEATS66H       *Figure 4. Final seasonally adjusted seris +)|   Figure 5. Final trend.   !Figure 6. Final irregular factors""V    #Figure 7. Forecasts by both methods$$p     To take advantages both from X-12-ARIMA and TRAMO/SEATS researchers from CENSUS Bureau are developing X-13-ARIMA, which would integrate the best from empirically based method and method based one.         We have used STAMP program which uses structural time series modelling. Series = trend + seasonal + intervention + irregular All these components could be handled in several different ways. The results were satisfactory.           Conclusion After trying several forecasting and decomposition methods for tourism in Croatia we conclude that method TRAMO/SEATS is sligtly better when it comes to handling interventions in time series.       0` 33` Sf3f` 33g` f` www3PP` ZXdbmo` \ғ3y`Ӣ` 3f3ff` 3f3FKf` hk]wwwfܹ` ff>>\`Y{ff` R>&- {p_/̴>?" dd@,|?" dd@   " @ ` n?" dd@   @@``PR    @ ` ` p>> f(    6  `}  T Click to edit Master title style! !  0ط  `  RClick to edit Master text styles Second level Third level Fourth level Fifth level!     S  0H ^ `  >*  0<ę ^   @*  0Ǚ ^ `  @*H  0޽h ? 3380___PPT10.` Default Design 0 zrP (    00a P   a P*    0a    a R*  d  c $ ?  a  0a  0 a RClick to edit Master text styles Second level Third level Fourth level Fifth level!     S  6a _P  a P*    6a _  a R*  H  0޽h ? 3380___PPT10.0$$ 0  $(  r  S Lz L r  S L `   L H  0޽h ? 3380___PPT10.@>$  0 0 $(   r  S L `}  L r  S L ` L H  0޽h ? 3380___PPT10.@ E$  0 @$(  r  S pR `}  R r  S R ` R H  0޽h ? 3380___PPT10.y9`  0 p`(  r  S R `}  R r  S XR ` R 4  # A*9LH  0޽h ? 3380___PPT10.$a4  0  4(     BA   Rr  S >R `}  R H  0޽h ? 3380___PPT10. &4  0 $4(  $r $ S  R `}  R  $ BA   RH $ 0޽h ? 3380___PPT10.@\o!L 0 KKgtv!K(  tJ p sv# #"rf ` R u s *1 ?ppZB ?u s *1 ?pZB @u s *1 ?pZB Cu s *1 ?ZB Eu s *1 ?((pZB Hu s *1 ? pZB Lu s *1 ?44ZB Yu s *1 ?ZB fu s *1 ?hhZB su s *1 ?ZB u s *1 ?ZB u s *1 ?66ZB u s *1 ?ZB u s *1 ?jjZB u s *1 ?ZB u s *1 ?ZB u s *1 ?88ZB u s *1 ?ZB u s *1 ?llZB u s *1 ?  ZB v s *1 ?  ZB v s *1 ?: : ZB v s *1 ?  ZB )v s *1 ?n n ZB 6v s *1 ?  ZB Cv s *1 ?  ZB Pv s *1 ?< < ZB ]v s *1 ?  H t 0޽h ? 3380___PPT10.=h5  0 44KHh4(  04  :  F# #"VJ ` W j Bt|W?Y :  V0.15 [0, 1] ad-hoc(   i B oW?e Y :  F--(   h B`?W? e:  e!Combined statistic Q (M1, M3-M11)(" !  g BW?Y   E(   f BW?e Y  E(   e BW? e  ^CRITERIA FOR DECOMPOSITION(   d B8W?Y l   [2.40% [0%, 5.0%] ad-hoc(   c BW?elY   [3.59% [0%, 5.0%] ad-hoc(   b BW?le  ZPercentage of outliers(   a BW?Y  l E(   ` BW?eY l E(   _ BlW?el LOUTLIERS(    ^ BW?Y 8  ] 6.26% [0%, 15.0%] ad-hoc(   ] BW?e8Y  F--(   \ BdW?8e aForecast error over last year(   [ BW?Y  8 E(   Z BW?eY 8 E(   Y BxW?e8 RFORECAST ERROR(   X BW?Y   ] 4.84 [1.75, 4.25] 0.1%*   W BW?eY  [ 3.80 [2.21, 3.79] 5%*   V BDY?e jKurtosis (significant)8     U BY?Y j  F--(   T B Y?ejY  Z 0.12 [-0.40, 0.40] 5%(   S B@Y?je fSkewness(    R BY?Y  j F--(   Q BTY?eY j V 4.29 [0, 5.99] 5%(   P BY?ej M Normality(     O BY?Y 6  E(   N BP"Y?e6Y  E(   M B&Y?6e \DESCRIPTION OF RESIDUALS(   L BL+Y?Y  6 F--(   K BY?eY 6 V 0.02 [0, 5.99] 5%(   J Bl3Y?e6 cBox-Pierce on squared residuals(    I B8Y?Y   R-- [0, ?] 0.1%(   H B BhY?4e [STATISTICS ON RESIDUALS(   = BmY?Y 4 X5.632 [0, 10] ad-hoc(   < BrY?eY 4 X3.554 [0, 10] ad-hoc(   ; BvY?e4 cSA quality index (stand. to 10)(    : B {Y?Y  ZModel 2 (X-12-Arima)*   9 BY?eY  }Model 1 (Tramo-Seats)*    8 BY?e `Information on Diagnostics*  ZB k s *1 ? ZB l s *1 ?: : ZB m s *1 ?: ZB n s *1 ? : ZB q s *1 ? ZB s s *1 ?ee: ZB v s *1 ?Y Y : ZB z s *1 ?4 4ZB  s *1 ? ZB  s *1 ?h hZB  s *1 ? ZB  s *1 ? ZB  s *1 ?6 6ZB Ȉ s *1 ? ZB Ո s *1 ?j jZB  s *1 ? ZB  s *1 ? ZB  s *1 ?8 8ZB   s *1 ? ZB  s *1 ?l lZB # s *1 ?  ZB 0 s *1 ? H  0޽h ? 3380___PPT10.z4  0 4(  r  S Y `}  Y   BA Y3M YH  0޽h ? 3380___PPT10.`64  0 4(  r  S dY `}  Y   BA   YH  0޽h ? 3380___PPT10.*=4  0  4(  r  S H?Y `}  Y   BA   YH  0޽h ? 3380___PPT10.'J4  0 04(  r  S Y `}  Y   BA   YH  0޽h ? 3380___PPT10.9s  0 P(  r  S |ZZT}   H  0޽h ? 3380___PPT10.Z,  0 @(  r  S  ` L H  0޽h ? 3380___PPT10.PN$  0 `$(  r  S  `}   r  S ) `  H  0޽h ? 3380___PPT10.[rP`2aBDF_<HMKMO^N12( / 0LDArial 00TTsܖ 0ܖ B .  @n?" dd@  @@`` P)3. ResultsCFigure 2. Seasonal factors extracted by X-12-ARIMA and TRAMO/SEATS6Figure 3. Final trend with X-12-ARIMA and TRAMO/SEATSSlide 7Slide 8+Figure 4. Final seasonally adjusted seris Figure 5. Final trend"Figure 6. Final irregular factors$Figure 7. Forecasts by both methods Slide 13 Slide 14 Conclusion  Fonts UsedDesign Template Slide Titles-$xx"_} 0Ante RozgaAnte Rozga-.  2 q1X."System9-@Arial-. 2 * Forecasting.-@Arial-.  2 *Gwith.-.  @n?" dd@  @@`` P)   qu I"#$%&o"$>2I]Tto;p"$:=>0bO p"$;i wU^}D `"$oe16V{Yg !"$]XX%DxG /"$~iljM"n< 0AA@3ʚ;ʚ;g4OdOd 0ppp@ <4dddd w 0Ts 80___PPT10 pp? ,O  =U-1Forecasting with Intervention: Tourism in Croatia< !Ante Rozga, Toni Marasovi, Josip Arneri University of Split, Croatia"GZ =F1. IntroductionTourism is among the most vulnerable business activities. It could be affected by political crisis, outbreak of the desease, economic crisis and war activities. In Croatia, the war for independence in 1991 affected tourism seriously. The number of foreign tourist has fallen more than 85% in 1992, compared with 1989. But, there were another interventions: the military action  Storm in August 1995 for deliberation of occupied Croatian teritories and NATO strike in 1999 against Serbia, connected with Kosovo crisis. Although NATO action was not conducted on Croatian teritory, the action had impact on Croatian tourism. rrO            2. Methods"bWe have used several satistical methods to analyze seasonal and other variations in monthly time series. Some of them are empirically based while the others were models based methods. We compared their performance to see the difference. We concentrated mostly on three of them: 2.1. X-12-ARIMA, developed by the Census Bureau, U.S.A. It is empirically based method ( ad-hoc method ), still dominant method for seasonal adjustment throughout the world. 2.2. TRAMO/SEATS, developed in Banco de Espaa, Madrid, by Gomes, Maravall and Caporello. This method is popular in EU. 2.3. Structural Time Series Model developed by Harvey and others, computer programe by Timberlake Consultancy Inc.                     3. Results"eWe have analyzed nights spent by tourists from July 1993 until April 2007. Figure 1. Nights in 000 ff&  BFigure 2. Seasonal factors extracted by X-12-ARIMA and TRAMO/SEATSCC|    5Figure 3. Final trend with X-12-ARIMA and TRAMO/SEATS66H       *Figure 4. Final seasonally adjusted seris +)|   Figure 5. Final trend.   !Figure 6. Final irregular factors""V    #Figure 7. Forecasts by both methods$$p     To take advantages both from X-12-ARIMA and TRAMO/SEATS researchers from CENSUS Bureau are developing hybrid X-13-ARIMA-SEATS, which would integrate the best from empirically based method and method based one.        We have used STAMP program which uses structural time series modelling. Series = trend + seasonal + intervention + irregular All these components could be handled in sev  !"$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~Root EntrydO)Р`@PicturesDCurrent User8>SummaryInformation(4 PowerPoint Document(#}DocumentSummaryInformation8)   qu I"#$%&o"$>2I]Tto;p"$:=>0bO p"$;i wU^}D `"$oe16V{Yg !"$]XX%DxG /"$~iljM"n< 0AA@3ʚ;ʚ;g4OdOd 0ppp@ <4dddd w 0Ts 80___PPT10 pp? ,O  =?-1Forecasting with Intervention: Tourism in Croatia< !Ante Rozga, Toni Marasovi, Josip Arneri University of Split, Croatia"GZ =F1. IntroductionTourism is among the most vulnerable business activities. It could be affected by political crisis, outbreak of the desease, economic crisis and war activities. In Croatia, the war for independence in 1991 affected tourism seriously. The number of foreign tourist has fallen more than 85% in 1992, compared with 1989. But, there were another interventions: the military action  Storm in August 1995 for deliberation of occupied Croatian teritories and NATO strike in 1999 against Serbia, connected with Kosovo crisis. Although NATO action was not conducted on Croatian teritory, the action had impact on Croatian tourism. rrO            2. Methods"LWe have used several satistical methods to analyze seasonal and other variations in monthly time series. Some of them are empirically based while the others were models based methods. We compared their performance to see the difference. We concentrated mostly on three of them: 2.1. X-12-ARIMA, developed by the Census Bureau, U.S.A. It is empirically based method ( ad-hoc method ), still dominant method for seasonal adjustment throughout the world. 2.2. TRAMO/SEATS, developed in Banco de Espaa, Madrid, by Gomes and Maravall. This method is popular in EU. 2.3. Structural Time Series Model developed by Harvey and others, computer programe by Timberlake Consultancy Inc.l                    3. Results"eWe have analyzed nights spent by tourists from July 1993 until April 2007. Figure 1. Nights in 000 ff&  BFigure 2. Seasonal factors extracted by X-12-ARIMA and TRAMO/SEATSCC|    5Figure 3. Final trend with X-12-ARIMA and TRAMO/SEATS66H       *Figure 4. Final seasonally adjusted seris +)|   Figure 5. Final trend.   !Figure 6. Final irregular factors""V    #Figure 7. Forecasts by both methods$$p     To take advantages both from X-12-ARIMA and TRAMO/SEATS researchers from CENSUS Bureau are developing hybrid X-13-ARIMA-SEATS, which would integrate the best from empirically based method and method based one. 6       We have used STAMP program which uses structural time series modelling. Series = trend + seasonal + intervention + irregular All these components could be handled in several different ways. The results were satisfactory.           Conclusion After trying several forecasting and decomposition methods for tourism in Croatia we conclude that method TRAMO/SEATS is sligtly better when it comes to handling interventions in time series.      0 P(  r  S pTZT}  T H  0޽h ? 3380___PPT10.Z,rZN 613( / 0LDArial 00TTsܖ 0ܖ B   !"#$%&'(*+,-./01234567Oh+'0 `h    (4Forecasting with Intervention: Tourism in Croatia Ante Rozga Ante Rozga18Microsoft Office PowerPoint@ {!@"@H`Gg  b%  y--$xx--'@Arial-.  2 q1X."System9-@Arial-. 2 * Forecasting.-@Arial-.  2 *Gwith.-@Arial-. 2 *[Intervention: .-@Arial-. "2 6(Tourism in Croatia .-@Arial-. 2 JA Ante Rozga.-@Arial-.  2 J^,X.-@Arial-. 2 Q- Toni Marasovi.-@Arial-.  2 QNcX.-@Arial-. 2 QQ, Josip Arneri.-@Arial-.  2 QpcX.-@Arial-. %2 X;University of Split,.-@Arial-. 2 _HCroatia.-՜.+,0X    On-screen ShowHOME} ArialDefault Design2Forecasting with Intervention: Tourism in Croatia1. Introduction 2. Methods eral different ways. The results were satisfactory.           Conclusion After trying several forecasting and decomposition methods for tourism in Croatia we conclude that method TRAMO/SEATS is sligtly better when it comes to handling interventions in time series.    $  0 @$(  r  S TG `}  G r  S UG ` G H  0޽h ? 3380___PPT10.y9r<FHrJ13( / 0LDArial 00TTsܖ 0ܖ B .  @n?" dd@  @@`` P)   qu I"#$%&o"$>2I]Tto;p"$:=>0bO p"$;i wU^}D `"$oe16V{Yg !"$]XX%DxG /"$~iljM"n< 0AA@3ʚ;ʚ;g4OdOd 0ppp@ <4dddd w 0Ts 80___PPT10 pp? ,O  =U-1Forecasting with Intervention: Tourism in Croatia< !Ante Rozga, Toni Marasovi, Josip Arneri University of Split, Croatia"GZ =F1. IntroductionTourism is among the most vulnerable business activities. It could be affected by political crisis, outbreak of the desease, economic crisis and war activities. In Croatia, the war for independence in 1991 affected tourism seriously. The number of foreign tourist has fallen more than 85% in 1992, compared with 1989. But, there were another interventions: the military action  Storm in August 1995 for deliberation of occupied Croatian teritories and NATO strike in 1999 against Serbia, connected with Kosovo crisis. Although NATO action was not conducted on Croatian teritory, the action had impact on Croatian tourism. rrO            2. Methods"bWe have used several satistical methods to analyze seasonal and other variations in monthly time series. Some of them are empirically based while the others were models based methods. We compared their performance to see the difference. We concentrated mostly on three of them: 2.1. X-12-ARIMA, developed by the Census Bureau, U.S.A. It is empirically based method ( ad-hoc method ), still dominant method for seasonal adjustment throughout the world. 2.2. TRAMO/SEATS, developed in Banco de Espaa, Madrid, by Gomes, Maravall and Caporello. This method is popular in EU. 2.3. Structural Time Series Model developed by Harvey and others, computer programe by Timberlake Consultancy Inc.                     3. Results"eWe have analyzed nights spent by tourists from July 1993 until April 2007. Figure 1. Nights in 000 ff&  BFigure 2. Seasonal factors extracted by X-12-ARIMA and TRAMO/SEATSCC|    5Figure 3. Final trend with X-12-ARIMA and TRAMO/SEATS66H       *Figure 4. Final seasonally adjusted seris +)|   Figure 5. Final trend.   !Figure 6. Final irregular factors""V    #Figure 7. Forecasts by both methods$$p     To take advantages both from X-12-ARIMA and TRAMO/SEATS researchers from CENSUS Bureau are developing hybrid X-13-ARIMA-SEATS, which would integrate the best from empirically based method and method based one.        We have used STAMP program which uses structural time series modelling. Series = trend + seasonal + intervention + irregular All these components could be handled in several different ways. The results were satisfactory.           Conclusion After trying several forecasting and decomposition methods for tourism in Croatia we conclude that method TRAMO/SEATS is sligtly better when it comes to handling interventions in time series.    rJJ}1Root EntrydO)P hS @PicturesDCurrent User8/SummaryInformation(4   !"#$%&'(*+,-./01234567_}rozgarozgae Rozga-.  2 q1X."System9-@Arial-. 2 * Forecasting.-@Arial-.  2 *Gwith.-