An overview on XML similarity: background, current trends and future directions

Size: px
Start display at page:

Download "An overview on XML similarity: background, current trends and future directions"

Transcription

1 Sumitt to Elsvi Sin An ovviw on XML simility: kgoun, unt tns n futu itions Jo Tkli, Rih Chi *, n Kokou Ytongnon LE2I Lotoy UMR-CNRS, Univsity of Bougogn, Dijon Cx Fn ABSTRACT In nt ys, XML hs n stlish s mjo mns fo infomtion mngmnt, n hs n oly utiliz fo omplx t psnttion (.g. multimi ojts). Owing to n unplll insing us of th XML stn, vloping ffiint thniqus fo omping XML-s oumnts oms ssntil in th ts n infomtion tivl ommunitis. In this pp, w povi n ovviw of XML simility/ompison y psnting xisting sh lt to XML simility. W lso til th possil pplitions of XML ompison posss in vious fils, nging ov t whousing, t intgtion, lssifition/lusting n XML quying, n isuss som qui n mgnt futu sh itions Elsvi Sin. All ights sv. Kywos: XML, Smi-stutu t; Stutul simility; T it istn; Dt whousing; Doumnt lssifition n lusting; Infomtion tivl; Rnk quis. Contnts 1. Intoution A Glimps on XML Doumnt-nti Vs Dt-nti XML XML Dt Mol Bkgoun T Eit Distn Mthos fo XML Simility Bsi Notions n Conpts Stt of th At in T Eit Distn Mthos Infomtion Rtivl Mthos fo XML Simility Titionl Infomtion Rtivl Extning Convntionl Infomtion Rtivl to Dl with XML Oth thniqus fo XML Simility Stutu-only XML Simility Mthos Stutu-n-ontnt XML Simility Mthos Applitions of XML Simility Dt Whousing: Vsion Contol n Chng Mngmnt XML Clssifition n Clusting Dt Intgtion Rnk XML Quying Disussions n futu sh itions XML Stutul Simility Untt Su-t Similitis Th spil s of singl lf no su-ts XML Smnti Simility Exploiting XML Gmms Conlusion * Cosponing utho. Tl.: ; Fx: ; -mil: ih.hi@u-ougogn.f

2 2 Sumitt to Elsvi Sin 1. Intoution W3C s XML (Xtnsil Mk-up Lngug) hs ntly gin unplll impotn s funmntl stn fo ffiint t mngmnt n xhng. Infomtion stin to ost ov th w is hnfoth psnt using XML, in o to gunt its intopility. Th us of XML ovs t psnttion n stog, ts infomtion inthng, t filting, s wll s w svis inttion. Owing to th insing w xploittion of XML, XML-s simility/ompison oms ntl issu in th ts n infomtion tivl ommunitis. By XML simility, w unlin XML oumnt-lt similitis, i.., oumnt/oumnt, oumnt/pttn 1, s wll s oumnt/gmm 2 ompison, th typs of ojts ing omp vying w..t. (with spt to) th pplition snio t hn. Applitions of XML ompison numous n ng ov: i) vsion ontol, hng mngmnt n t whousing (fining, soing n owsing hngs twn iffnt vsions of oumnt, suppot of tmpol quis n inx mintnn), ii) smi-stutu t intgtion (intifying simil XML oumnts oiginting fom iffnt t sous, to intgt so tht th us n ss mo omplt infomtion), iii) lssifition/lusting of XML oumnts gth fom th w ginst st of XML gmms l in n XML ts (just s shms nssy in titionl DBMS fo th povision of ffiint stog, tivl n inxing filitis, th sm is tu fo XML positois), iv) s wll s XML tivl (fining n nking sults oing to thi simility in o to tiv th st sults possil). A wi ng of lgoithms fo omping smistutu t,.g., XML-s oumnts, hv n popos in th littu. Ths vy w..t. th kins of XML t thy onsi, s wll s th kins of pplitions thy pfom. Thy n lssifi in th min goups: i) Eit Distn (ED) s mthos, ii) Infomtion Rtivl (IR) s mthos, n iii) oth ivs ppohs xploiting iffnt thniqus to XML ompison (.g., g mthing, pth simility ). EDs mthos mk us of ynmi pogmming thniqus fo fining th it istn twn t stutus, XML oumnts ing mol s o ll ts. Most of ths mthos sign fo oumnt/oumnt ompison tsks. Thy tgt igoously stutu XML n usully fin-gin. Thy minly usful fo pplitions tht qui ut (fin-gin) ttion of XML oumnt stutul similitis, i.., vsion ontol n hng mngmnt (ED lgoithms hving th vntg of 1 An XML pttn is potion of n XML oumnt. 2 Eith DTD o XML Shm [28]. pouing it sipts, long th simility vlu itslf, whih woul xploit in siing hngs), t intgtion, s wll s XML lssifition/lusting pplitions. IR-s ppohs xtn onvntionl infomtion tivl mthos,.g., th vto sp mol, so s to povi XML oumnt/quy simility ssssmnt. In this ontxt, n XML quy silly oms own to ith n XML oumnt, pttn, o onjuntion of pttns (f. Stion 3.2). IR-mthos tgt loosly stutu XML t n usully osgin, thus usful n gnlly xploit fo fst simpl XML sh n tivl. Not tht nk XML quying pplitions usully pioitiz pfomn w..t. sult qulity, i.., pouing goo-nough sults in shot tim lps. Th thi goup of mthos ompiss of iffnt ppohs to XML simility. Thy xploit vious thniqus (.g., tg simility, g mthing, pth simility, ntopy ) ssing spifi pplition snios. Som povi ppoximtions of (mo omplx n ut) xisting ppohs (minly ED-s). Th gol of this stuy is to povi unifi viw of th polm, ssssing th iffnt spts, thniqus n vious pplitions lt to XML simility. Th min of this pp is ogniz s follows. Stion 2 psnts glimps on XML s t psnttion mol. Stion 3 viws kgoun in XML simility, oving th th min goups of mthos mntion ov: EDs, IR-s n vious pplition spifi ppohs. Stion 4 vlops th min pplitions n uss of XML ompison. Som ongoing motivtions n possil futu sh itions ov in th Stion 5. Stion 6 onlus th pp. 2. A Glimps on XML With th gowth of th Wol Wi W, th is n insing n to utomtilly poss W oumnts fo ffiint t mngmnt, simility lusting n sh pplitions. Whil HTML (Hyp Txt Mkup Lngug) povis th visul mkup, hving knowlg of th logil stutu of th t is funmntl pquisit fo th intopility of ws infomtion systms [31]. Hn, XML ws intou y th W3C s n ffiint mns fo t psnttion n mngmnt Doumnt-nti Vs Dt-nti XML Th two iffnt viws of XML: th oumntnti viw n th t-nti viw. Doumnt-nti sign involvs lil us of f-fom txt tht is mk up with lmnts (f. Figu 1.). It fouss on XML pplitions fo xhnging oumnts in th titionl sns. Doumnt-nti XML is sy to n on som

3 Sumitt to Elsvi Sin 3 sot of output vi. Fo xmpl, it is quit sy to fomt th oumnt in Figu 1.. into HTML, us it is simil to HTML in its tgging n mk up styls. Dt-nti XML, howv, unlins wll stutu infomtion, i.. th t is stitly tgg (f. Figu 1.). Th t-nti viw is usully utiliz fo xhnging t in stutu fom, suh s lssil EDI (Eltoni Dt Inthng) n ts pplitions. In ition, t-nti oumnts si to poss with omput pogms n utomti posss, us th t is tt ogniz. <?XML> <Mmo> Pls mk su you t th <Lotion>ninth floo</lotion>y <MtingTim>10:30 AM</MtingTim> to<pupos>isuss th ugt</pupos> </Mmo>. Doumnt-nti XML <?XML> <Mmo> <MtingTim>10:30AM</MtingTim> <Pupos>Disuss Bugt</Pupos> <Lotion>ninth floo</lotion> </Mmo>. Dt-nti XML Fig. 1 - Smpl oumnt-nti n t-nti XML oumnts. Both t-nti n oumnt-nti spts of XML hv n onsi in ssssing XML simility. As mntion pviously, ED-s XML simility ppohs usully fin-gin n thus it to omping wll stutu t-nti XML oumnts, whil IR-s mthos os-gin n tgt loosly stutu oumnt-nti XML XML t mol XML oumnts psnt hihilly stutu infomtion n gnlly mol s O Ll Ts (OLTs) 1. In titionl DOM (Doumnt Ojt Mol) o ll t [88], nos psnt XML lmnts n ll with osponing lmnt tg nms. Elmnt ttiuts mk th nos of thi ontining lmnts. Som stuis hv onsi OLTs with istint ttiut nos, ll with osponing ttiut nms [61], [90]. Attiut nos pp s hiln of thi nompssing lmnt nos, sot y ttiut nm, n pping fo ll su-lmnt silings [61]. Elmnt/ttiut vlus n isg (stutuonly) o onsi (stutu-n-ontnt) in th ompison poss following th pplition snio t hn (f. Figu 2). In gnl, lmnt/ttiut vlus isg whn vluting th stutul poptis of htognous XML oumnts, i.., oumnts oiginting fom iffnt t-sous n not onfoming to th sm gmm (DTD/XML Shm) 2, so s to pfom XML stutul lssifition/lusting [23] [61] o stutul quying (i.., quying th stutu of oumnts, isging ontnt [6]). Nonthlss, vlus usully tkn into ount with mthos it to XML hng mngmnt [16] [20], t intgtion [37] [51], n XML (stutu-n-ontnt) quying pplitions [74], [90], wh oumnts tn to hv ltivly simil stutus (poly onfoming to th sm gmm [47], [86]). With suh mthos, XML txt squns usully ompos into wos, mpping h wo to lf no ll with th sptiv wo. <?XML> <Amy> <Fulty> <Dptmnt> <Pofsso>John Cm </Pofsso> <Stunt>John Tkgi </Stunt> </Dptmnt> </Fulty> </Amy>. XML oumnt 4 1 Amy 2 3 Pofsso Fulty Dptmnt Stunt John Cm John Tkgi OLT with vlus. OLTs wh vlus isg Fig. 2. A smpl XML oumnt with osponing OLTs. 3. Bkgoun on XML Simility 1 Amy Pofsso Stunt Vious itions oul llow th siption n tgoiztion of XML simility mthos, inluing: i. Th kin of thniqu ing us: ED-s, IRs, n oths (tg simility, g mthing, ) ii. Th kin of XML t ing omp: oumnt/oumnt, oumnt/pttn o oumnt/gmm oumnt-nti o t-nti, stutu-only o stutu-n-ontnt iii. Th intn pplition omin: hng mngmnt n vsion ontol, t intgtion, lssifition/lusting, n nk quying Fulty Dptmnt 5 1 In th following, t signts o ll t. 2 It is th s of lots of XML oumnts on th w [61].

4 4 Sumitt to Elsvi Sin In th following, fo lity of psnttion, w viw XML simility mthos s on th kins of thniqus thy xploit (i.., ED-s, IR-s n oths). W stimt this tgoiztion povis th simplst n most onsistnt unifi viw of th wi vity of ivs mthos popos in th littu. Th kins of XML t ing tt s wll s th intn pplitions omins will isuss fo h mtho. Ctlogs summizing th poptis n htistis of ll mthos ov in this viw pit in Tls 1, 2 n 3 t th n of th pp T it istn mthos fo XML simility Vious mthos, fo stimting th similitis twn hihilly stutu t, ptiully twn XML oumnts, hv n popos in th littu. Most of thm iv, in on wy o noth, th ynmi pogmming thniqus fo fining th it istn twn stings [48], [83], [87]. In ssn, ll ths ppohs im t fining th hpst squn of it options tht n tnsfom on t into noth, XML oumnts ing moll s o lll ts [88]. Blow, w povi foml finitions of th ommon onpts lt to t ED Bsi notions n ommon onpts Dfinition 1 - Eit sipt: It is squn of it options op 1, op 2,, op k. Whn ppli to t T, th sulting t T is otin y pplying it options of th Eit Sipt (ES) to T, following thi o of ppn in th sipt. By ssoiting osts with h it option, Cost Op, th ost of n ES is fin s th sum of ES th osts of its omponnt options: Cost ES = Cost. i=1 Op Dfinition 2 - Eit istn: Th it istn twn two ts A n B is fin s th minimum ost of ll it sipts tht tnsfoms A to B: Dist(A, B) = Min{Cost ES }. Thus, th polm of omping two ts A n B, i.. vluting th stutul simility twn A n B, is fin s th polm of omputing th osponing t it istn [89]. As fo t it options, thy iff following th it istn mtho t hn, n n lssifi in two goups: tomi t it options n omplx it options. An tomi it option on t (i.. oot o ll t) is ith th ltion of n inn/lf no, th instion of n inn/lf no, o th plmnt (i.. upt) of no y noth on. A omplx t it option is st of tomi t it options, tt s on singl option. A omplx t it option is ith th instion of whol t s su-t in noth t (whih is tully squn of tomi no instion options), th ltion of whol t (whih onsists of squn of tomi no ltion options), o moving su-t fo on position into noth in its ontining t i (whih is squn of tomi no instion/ltion options). Th uthos in [17] intou su-t opying n gluing options. Ths simil to t instions/ltions sptivly, ut fin in th ontxt of uno t ompison. Thus, thy won t futh invstigt hun. In th following, w psnt th gnl fom n vitions fo h of th t it options stt ov. Dfinition 3 Upt no: Givn no x in t T n nw no y, Up(x, y) is n upt option pling x y y in th tnsfom t (f. Figu 3.). No y will hv th sm pnt n hiln s x. Dfinition 4 Inst no: Vition 1: Givn no x n t T, T nompssing no p with fist lvl su-ts (i.. hiln) P 1,, P m, Ins(x, p, {P i,, P j }) insts no x in T s th i th hil of p mking x th pnt of th onsutiv susqun of suts {P i,, P j } of p (f., Figu 3.). Vition 2: Th instion option n stit to lf nos in som ED ppohs: InsLf(x, p, i) insting lf no x s th i th hil of p (f., Figu 3.) f. Upt no. Inst (intnl) no. Dlt (intnl) no. Inst lf no. Dlt lf no f. Mov no (su-t) InsT(S,, 1) Up(, f) f DlLf(, ) f Dl(f, ) S S DlT(S, ) g. Inst su-t h. Dlt su-t Fig. 3 - T it options (simplifi syntxs). f f InsLf(,, 2) Ins(f,, {, }) Mov(f,, 2) f f

5 Sumitt to Elsvi Sin 5 Dfinition 5 Dlt no: Vition 1: Givn no x n t T, T nompssing no p with fist lvl su-ts (i.. hiln) {P 1,, P i-1, x, P i+1,, P m }, n x hving fist lvl su-ts {X 1,, X n }, Dl(x, p) lts no x in T, mking th hiln of x om th hiln of p. Th hiln of x inst in th pl of x s susqun in th lft-to-ight o of th hiln of p (Figu 3.). Vition 2: Th ltion option n stit to lf nos in som ED mthos: DlLf(x, p) (f., Figu 3.). Dfinition 6 - Inst t: Givn t A n t T, T inluing no p with fist lvl su-ts {P 1,, P m }, InsT(A, p, i) is t instion ppli to T, insting A s th i th su-t of p (f. Figu 3.g). Dfinition 7 - Dlt t: Givn t A n t T, T nompssing no p with fist lvl su-ts {P 1,, P i-1, A, P i+1,, P n }, DlT(A, p) lts su-t A in T fom mong th hiln of p (f. Figu 3.h). Dfinition 8 Mov no: Givn no x n t T, T nompssing no p, Mov(x, p, i) movs x n its hiln to om th i th hil of no p (f. Figu 3.f) Stt of th t in t it istn mthos T ED lgoithms n istinguish y th st of it options tht llow s wll s ovll omplxity/pfomn n optimlity/ffiiny lvl. Ely ppohs: In [78], th utho intous th fist non-xponntil lgoithm to omput th it istn twn o ll ts, llowing instion, ltion n sustitution (lling) of inn nos n lf nos. Th sulting lgoithm hs omplxity of O( T 1 T 2 pth(t 1 ) 2 pth(t 2 ) 2 ) whn fining th minimum it istn twn two ts T 1 n T 2 ( T 1 n T 2 not t inlitis whil pth(t 1 ) n pth(t 2 ) th pths of th ts). Similly, ly ppohs in [77] n [89] llow instion, ltion n lling of nos nywh in th t (f. Figus 3., 3. n 3.). Yt, thy min omplx. Th ppoh in [89] is of omplxity O( 2 T 1 min( T 1, T 2 )) ( is th wight it istn 1 ), whil tht in [77] hs tim omplxity of O( T 1 T 2 pth(t 1 ) pth(t 2 )). Not tht th ppohs in [77], [78], [89] w not minly vlop in th XML ontxt, n thus might yil sults (i.. it sipts, n onsquntly istns) tht not ompltly ppopit to XML t. In ptiul, it hs n ntly gu tht stiting instion n ltion options to lf nos fits tt in th ontxt of XML t, w..t. instions n ltions ppli nywh in th XML t 1 Lt S = op1, op 2,, op n th hpst squn of it options tht tnsfoms t A to B, thn th wight it istn is givn y = 1 i n w i wh w i, fo 1 i n, is 1 if op i is n inst o lt option, n 0 othwis. [23]. Following [23], th ltt pi of options stoys th mmship stitions of th XML hihy n thus os not sm to ntul fo XML t (.g., lting/insting n inn no n moving its hiln up/own on lvl, f. Figus 3. n 3.). ED-s ppohs it to XML-s t tn to pvnt suh options y utilizing ons tht tgt lf nos (f. Figus 3. n 3.) o whol su-ts (f. Figus 3.g n 3.h). Hn, th ltion of n intnl no in n XML oumnt t woul qui ltions of ll nos in its pth, stting fom th lf no n going up to th intnl no itslf, whih oul lso pfom vi on singl t ltion option. Likwis, th instion of n inn no must pfom fo th instion of ny of its snnts, whih oul untkn vi on t instion option. Ting qulity fo pfomn: In [16], [20], th uthos stit instion n ltion options to lf nos n mov opto tht n lot su-t, s singl it option, fom on pnt to noth (f. Figus 3., 3., 3. n 3.f). Not tht mov option n sn s sussion of inst n lt options. Yt, it is iffnt in its ost (th uthos in [16], [20] onsi tht th ost of moving su-t is muh lss thn th sum of th osts of lting th sm su-t, no y no, fom its unt position n insting it in its nw lotion). Howv, lgoithms in [16], [20] o not gunty optiml sults. On on hn, th lgoithm in [16] uns in fiv phss: upt, lign, inst, mov n lt. Eh of thm, to th xption of th lign phs, ospons to th pplition of th st possil omintions of th osponing it option, so s to mth th omp ts. As fo th lign phs, th uthos mk us of vition of th Longst Common Susqun lgoithms (LCS) [59] in tmining th st of mislign hiln nos 2 of givn inn no (pt fo h t no thoughout th mthing poss). Thos mislignmnts xploit in th susqunt mov phs in tmining th st of mov options to pfom. Nonthlss, XML oumnts ing omp shoul i two min ssumptions without whih th lgoithm woul yil suoptiml sults (i.., ovlooking th minimum ost it sipt): i) no lls hv to follow tin pfin oing w..t. givn shm, n ii) givn ny lf no in th fist oumnt, th is t most on lf no in th son oumnt tht mths th fist n vi vs. Algoithm omplxity simplifis to O(n + 2 ), wh n is th totl num of lf nos in th ts ing omp. On th oth hn, th lgoithm in [20] tis to tt lg su-ts tht w lft unhng twn th two XML ts ing omp. Ths mth. Consquntly, it tis to mth mo nos y onsiing nstos n snnts of mth nos, tking lls 2 Ths nos ing th sm lls/vlus ut hving iffnt oing positions w..t. thi pnt nos.

6 6 Sumitt to Elsvi Sin into ount. It lso onsis ID ttiuts in mthing osponing nos (nos with intil IDs mth). Th lgoithm howv ts som qulity to gt n lgoithm whih uns in vg lin tim: no mo thn O(N log(n)) wh N is th mximum num of nos in th ts ing omp. In oth wos, th istn ttin whn omping two ts is not lwys miniml, som sts of mov options not ing optiml (i.., th pou it sipt is not of miniml ost). Both mthos in [16] n [20] w vlop in th ontxt of hng mngmnt n vsion ontol. Thy sign fo oumnt/oumnt ompison n onsi XML lmnt/ttiut vlus (f. Figu 2.) in thi omputtions, in ontst with mining mthos in this stion whih spifilly tgt th stutul poptis of XML oumnts (f. Figu 2.). Comining ffiiny n pfomn: Wok povi in [18] hs n onsi s fn point in nt t it istn littu n hs povi th sis fo vious XML lt stutul ompison stuis [61], [23], [79]. Chwth s ppoh stits instion n ltion options to lf nos (whih viw s ntul options in th XML ontxt) n llows th lling of nos nywh in th t, whil isging th mov option (f. Figus 3., 3. n 3.). Th popos lgoithm is it pplition of th fmous Wgn-Fish lgoithm [83] whih optimlity hs n it in o vity of omputtionl pplitions [2], [87]. It is lso mong th fstst t ED lgoithms vill. In shot, th utho tnsfoms ts into spil squns ll l-pis. Th l-pi psnttion of t oms own to th list, in po, of th l-pi psnttions of its nos. Th l-pi psnttion of t no is th pi (l, ) wh l is th no s ll n its pth in th t (.g., th l-pi psnttion of th XML t in Figu 2. is (Amy, 0), (Fulty, 1), (Dptmnt, 2), (Pofsso, 3), (Stunt, 3)). Consquntly, th utho simplifis th polm of omping two oumnt ts to tht of omping th osponing l-pi psnttions, using spiliztion of th Wgn-Fish lgoithm [83]. H lso xtns his lgoithm fo xtnl-mmoy omputtions n intifis sptiv I/O, RAM n CPU osts. Not tht this is th only lgoithm tht hs n xtn to ffiintly lult it istns in xtnl mmoy (without ny loss of omputtion qulity/ffiiny). Th ovll omplxity of Chwth s lgoithm is of O(N 2 ), its pfomn n ffiiny ing ogniz in [61] s wll s [23] (ll tht N is th mximum num of nos in th ts ing omp). Su-t simility: Whil it might onsi s stting point fo nt XML t it istn ppohs, th ppoh in [18] ovlooks tin su-t similitis whil omping XML oumnts. Fo instn, omputing it istn twn XML ts A, B n C in Figu 4 yils Dist(A, B) = Dist(A, C) = 3, whih ospons to th ost of th onsutiv inst options (unit osts usully us) intouing nos, n (, f n g) in t A tnsfoming it into B (C). Nonthlss, on n liz tht t A is stutully mo simil to B, thn to C, th su-t A 1, m up of nos, n, pping twi in B (B 1 n B 2 ) n only on in C (C 1 ). Suh sut stutul similitis lso lft unss y oth xisting ppohs. A 1 T B Fig. 4 - Smpl XML ts. In [61], th uthos stss th impotn of intifying su-t stutul similitis in n XML t ompison ontxt, u to th fqunt psn of pt n optionl lmnts in XML oumnts. Rpting lmnts oftn inu multipl ouns of simil lmnt/ttiut su-ts (psn of optionl lmnts/ttiuts) o intil su-ts in th sm XML oumnt (suh s th su-ts B 1 n B 2 in XML t B, f. Figu 4) whih flts th n to tk ths su-t smlns into onsition whil omping oumnts. Th uthos in [61] xtn th ppoh of Chwth [18] y ing two nw options: inst t n lt t (f. Figus 3.g n 3.h) to isov su-t similitis, y mking us of th ontin in ltion twn ts/su-ts. A t T 1 is si to ontin in t T 2 if ll nos of T 1 ou in T 2, with th sm pnt/ hil g ltionship n no o. Aitionl nos my ou in T 2 twn nos in th ming of T 1. Following [61], t A my inst in T only if A is ly ontin in th sou t T. Similly, t A my lt only if A is ly ontin in th stintion t T. Thfo, th popos ppoh ptus th su-t stutul similitis twn XML ts A/B in Figu 5, tnsfoming A to B in singl it option: (insting su-t B 2 in A, B 2 ouing in A s A 1 ), whs tnsfoming A to C woul lwys n th onsutiv inst options (insting nos, f n g). Th ovll omplxity of thi lgoithm simplifis to O(N 2 ). Stutul lusting xpimnts in [61] show tht th popos lgoithm outpfoms, in qulity, tht of Chwth [18], whih in tun yils tt sults thn Zhng n Shsh s lgoithm [89]. Howv, th uthos in [61] show tht thi lgoithm is onptully mo omplx thn its psso, quiing p-omputtion phs fo tmining th osts of t inst n lt options (whih omplxity is of O(2 N + N 2 )). T A f g B 1 B 2 C 1 C 2 T C

7 Sumitt to Elsvi Sin 7 Stutul summis: Th uthos in [23] povi n it istn lgoithm omining ftus fom oth [18], [61] n popos to pply it on XML t stutul summis, inst of whol oumnt ts, in o to gin in pfomn. Stutul summis pou using it ptition/nsting ution poss. Th stutul summy of n XML t oms own to moifi t in whih th unnis u to nstpt n pt XML nos limint (.g., th stutul summy of t B in Figu 4 is t A, th pt su-t B 2 ing omitt). Thi lgoithm n viw s spil s of [61] s lgoithm wh inst n lt t options osts omput s th sum of th osts of insting/lting ll iniviul nos in th onsi su-ts. Th lgoithm is of O(N 2 ) omplxity. Expimntl sults in [23] show impov oumnt lusting qulity w..t. [18] s lgoithm. XML/DTD simility: In [80], th uthos popos n ED-s ppoh tht is slightly iffnt in sop w..t. xisting mthos. Th uthos povi n lgoithm it to omping n XML oumnt n DTD gmm. Thy intou n lgoithm s on th t it istn onpt, s n fftiv n ffiint mns fo omping t stutus, XML oumnts n DTDs ing mol s o ll ts (lmnt/ttiut vlus in lmnts isg). Th popos mtho xtns th lgoithm in [61] y onsiing th vious DTD onstints on th xistn, ptility n ltntivnss of XML lmnts/ttiuts. Th ppoh is of polynomil omplxity (O(N 3 ) wh N is th mximum num of nos in th XML/DTD ts ing omp), in ompison with xisting xponntil mthos, i.., [6], psnt in th following stion. Clssifition xpimnts on sts of l n synthti XML oumnts, unlin th ppoh s fftivnss, n its ppliility to lg XML positois. Comping t it istn ppohs, to intify th st XML stutul simility mthos, is not tivil tsk. Eh mtho is vlop in spifi ontxt n might thus yil inppopit sults whn ppli in iffnt fmwok. Nonthlss, w..t. th two mjo itions tht oughly htiz t it istn mthos: ffiiny (qulity) n pfomn (omplxity), th ppoh in [61] sms to on of th most sophistit ED-s XML stutul simility mthos to t. Rging qulity, Nimn n Jgish s ppoh [61] ws povn to mo ut in tting XML stutul smlns, ptiully sut lt similitis, in ompison with som of its fmous pssos (i.. [18], [89]). It ws onsquntly opt s th sis fo mo nt stuis, ptiully [23], [80]. Rging pfomn, th uthos in [61] w l to mintin quti tim omplxity in vloping thi lgoithm, whih is typil to ED-s ppohs. Tl 1 pits th vious ED-s XML simility ppohs vlop in th littu Infomtion tivl mthos fo XML simility Whil lot of wok hs n untkn in it istn lt sh, fo omping XML t, XML simility is lso oming on of th ntl topis in th infomtion tivl fil. Rll tht ED-s ppohs fous on igoously stutu t-nti XML n tgt hng mngmnt, t intgtion n stutul lssifition/lusting pplitions, whs IRs mthos tt loosly stutu oumnt-nti XML n minly tgt nk XML quying. Sin oumnts flt with onvntionl infomtion tivl, i.. thy psnt unstutu t, titionl IR mthos fo shing n quying infomtion no long qut with smi-stutu t suh s XML [36]. As stt li, XML oumnts psnt hihilly stutu infomtion (f. Figu 2). In oth wos, ontnt is istiut t iffnt lvls of th oumnt t. Thfo, it is to tt iffntly w..t. flt ontnt so s to impov tivl pision [75]. Fist, infomtion pl n th oot no of n XML oumnt tns to mo impotnt thn infomtion futh own in th hihy [6] [90]. Intuitivly, s on sns in th XML t hihy, infomtion oms insingly spifi, onsisting of fin n fin tils, its fft on th mning of th whol oumnt t sing oingly. This is othogonl to titionl flt oumnts wh infomtion is pl t on singl lvl n is of th sm lvn. Son, uss in tun woul lik to f to oumnt stutu whn shing fo lvnt infomtion in XML oumnts [36]. To o so, thy pos so ll ontnt-nstutu quis, in ompison with th ontnt-only quis in onvntionl IR [69], y stiting th ontxt of intst to som XML lmnts in th oumnts ing sh [36]. Thfo, vious ttmpts to xtn xisting IR mthos in o to ount fo th stutu of XML oumnts, in th oumnt/quy ompison poss, hv n untkn ( quy, in th IR ontxt, unlining whol XML oumnt, n XML oumnt fgmnt o onjuntion of fgmnts). In th following sustion, w stt y psnting n ovll viw of titionl IR onpts, th vto sp mol in ptiul. Thft, w ov IR mthos xtn fo XML t Titionl infomtion tivl Infomtion tivl (IR) is nh of infomtis onn with th quisition, ogniztion, stog, sh n sltion of infomtion [69]. Whil it is us to l with flt txtul t (i.. lssil f txt oumnts), IR is ing xtn, sin th lst two s, so s to tt omplx infomtion suh s stutu/smi-stutu t (.g. XML, SGML, HTML, t.), imgs, gphis, souns n vios. In ssn, th gol of IR is to ffiintly intify/tiv, in t olltion, infomtion tht is

8 8 Sumitt to Elsvi Sin lvnt w..t. th us s ns [7]. Whil lvn in IR is o n impis notion, th stt onpt of lvn is gnlly ontiz y th notion of simility [65]. With onvntionl IR, oumnts n us quis usully onsist of sts of kywos. Thus, intifying oumnts tht lvnt (simil) to givn quy oms own to: Comping th kywos of h oumnt in th oumnt olltion to thos of th quy, Rnking th oumnts w..t. thi kywo similitis with th quy (oumnt sltion is untkn y fining simility thshol,.g. ng quis [1] o KNN quis [68]). Kywos ommonly wight in o to flt thi ltiv impotn in th quy/oumnt t hn. Th unlying i is tht tms tht of mo impotn in siing givn quy/oumnt ssign high wight. As wighting shm, th stn TF-IDF (Tm Fquny Invs Doumnt Fquny) ppoh (n its vints) of th vto sp mol [69] [70] is usully us. Not tht vious IR mols, oth thn th vto sp mol, hv n popos in th littu, mong whih th Booln mol [46], th poilisti mol [30], th LSI (Ltnt Smnti Inxing) mol [24], th DFR (Divgn Fom Rnomnss) mol [3], t. Howv, in this hpt, w stit ouslvs to th vto sp mol sin it is th most ommonly us, its pfomn ing it in o vity of pplitions n snios (.g., [19], [21], [27]). Most ttmpts to xtn IR thniqus, so s to ount fo XML t, fous on TF-IDF n th vto sp mol. Thfo, w ifly viw thos notions pio to isussing XML IR ppohs. In th stn vto sp mol, oumnts n quis inx in simil mnn, pouing vtos in sp whih imnsions psnt, h, istint inxing unit t i. An inxing unit usully stns fo singl tm, i.. kywo 1. Th ooint of givn oumnt D on imnsion t i, is not w D (t i ) n stns fo th wight of t i in oumnt D within oumnt olltion. w D (t i ) is omput using so of th TF-IDF fmily, tking into onsition oth oumnt n olltion sttistis. Consquntly, th lvn of oumnt D to quy Q, signt s Sim(Q, D), is vlut using msu of simility twn vtos suh s th inn pout, th osin msu, th J msu, th Di offiint, t., [7], [32], [52]. Fo instn, th osin msu, whih is on of th most ommonly us in th IR littu, is xpss s follows: Cos( u u Q, D ) = M =1 w (n ) w (n ) Q D M 2 n 2 w Q (n ) w D (n ) =1 =1 [ 0, 1] 1 A kywo n lso onsist of multipl wos (phs units). (1) As fo th TF-IDF so, iffnt vitions hv n popos [69], [70], [71]. W giv low th stn finition. Th TF-IDF so, mking up wight w D (t i ), ompiss of two ftos [69]: Th TF (Tm Fquny) fto whih signts th num of tims tm t i ous in oumnt D (oumnt sttistis). Th unlying i is tht th impotn of givn tm t i in siing oumnt D inss with th fqunt us of t i in D. Th IDF (Invs Doumnt Fquny) fto, mphsizing th ftion of oumnts tht ontin tm t i (olltion sttistis). Th unlying i is tht th impotn of givn tm t i in siing oumnt D ss with th fqunt us of t i in th oumnt olltion. A ommon TF-IDF mthmtil fomultion [70], [71] woul s follows. w D (t i ) = TF IDF hving: TF = tf(t i, D) unlining th num of tims tm t i ous in D N IDF = log f ( t, D) wh N is th totl num of i oumnts in th oumnt olltion, n f(t i, D) is th num of oumnts ontining tm t i Th TF-IDF so oul lso nomliz s til in [69], fo xmpl: N tf ( ti, D) log f ( ti ) wd ( ti ) = 2 N (2) t tf ( t, ) log i i D f ( ti ) Bus of its wll known ffiiny whn ling with flt txtul t, vious stuis hv fous on xtning th vto tivl TF-IDF s ppohs so s to tt smi-stutu t, i.. XML. W ov thos min ppohs in th following stion Extning onvntionl infomtion tivl to l with XML A num of thniqus xtning th vto sp mol tows fftiv XML infomtion tivl hv n sign, nmly [13], [31], [36], [75], [62]. Not tht vious oth thniqus,.g. [4], [56], hv lso n popos. Nonthlss, w limit ou psnttion h to th most si stuis, mining ppohs ing ov in th pplitions stion. Inxing nos: In [31], Fuh n Goßjohnn fin so-ll inxing nos. Ths tomi units in th XML oumnt, silly XML lmnts, whih nompss isjoint su-ts. Givn ths nos, TF-IDF wights n omput lolly, inst of ing

9 Sumitt to Elsvi Sin 9 vlut w..t. to th whol oumnt. Th wights woul thn ugmnt, own-wighting th sttistis (TF-IDF vlus) whn th tms popgt upws in th oumnt t, thus tking into ount th hihil spt of XML. Th unlying i is tht: th lg th istn twn no n its nsto, th lss tht no shoul ontiut to th lvn of its nsto s ontnt [31]. Following Fuh n Goßjohnn, th inxing nos to intifi xpliitly in it XML gmm finition (tht th uthos intify s n xtn DTD) osponing to th XML oumnt t hn (fo instn, th following hypothtil lmnt ltion <!ELEMENT Pofsso (#INDEX)> oul signt tht Pofsso lmnts in th XML oumnt of Figu 5 inxing nos). Fo xmpl, in Figu 5, nos Pofsso n Stunt onsi s inxing nos. Thus, TF-IDF sos omput fo ths nos s on thi txtul ontnt (.g. fo no Pofsso, sos fo John n Cm omput). Thn, ths sos popgt to upp nos in th oumnt t, multiplying thm y osponing ugmnttion wights (.g. fo no Pofsso, TF-IDF sos oughly multipli y fto of 0.6 whn popgt to no Dptmnt, thus sing thi lvn w..t. Dptmnt). Pofsso 0.6 Dptmnt 0.4 Stunt 0.5 John Cm John Tkgi 0.4 Jms Amy Fulty Stunt Tilo Fig. 5 - XML oumnt t with pfin ugmnttion wights. Not tht whil it mploys TF-IDF soing, th ppoh in [31] is lso uilt on onpts of th poilisti IR mol, using fts n uls to mp th XML ontnt n stutu. Nonthlss, it onstituts th uiling loks fo wll known vto tivl thniqu, i.. [36]. Singl/multi-tgoy tivl: In [36], Gs n Shk uil on Fuh n Goßjohnn s ppoh [31] y intouing th notion of tgoy. Following Gs n Shk, uss my wnt to f to givn pt (tgoy) of th XML oumnt in isoltion, qusting only ontnt tht is lvnt to tht tgoy, i.. singl tgoy tivl (.g. in Figu 5, th us might intst in pofssos tht only wok in lotois. Thus, only th ontnts of nos Pofsso osponing to tgoy Lotoy shoul onsi whn omputing tm Lotoy John 0.5 Pofsso Ptson sttistis. In oth wos, th so fo tm John osponing to Pofsso John Ptson shoul not fft y ouns of John in lmnts outsi th tgoy Lotoy). Likwis, in oth ss, uss my qust infomtion fom svl tgois, o thy o not to whih tgoy th qust ontnt longs, i.. multi-tgoy tivl (.g. in Figu 5, th us might intst in ll pofssos. Thus tm sttistis shoul ssss tking into ount ll osponing tgois). To solv th tgoy polm, th uthos popos to kp th inxs fo si nos (inxing nos [31]) n to iv qui inxs n sttistis fom th unlying si ons on-th-fly, i.. t quy untim. In oth wos, th vto sp is gnt ynmilly following us infomtion qusts. Tht is wht th uthos intify s flxil tivl, i.. th uss n ynmilly t quy tim fin th sop of thi quis. Inxing s on tm ontxt: In n ttmpt to futh ount fo th XML stutu in IR, Cml t l. [13] popos to xtn th vto sp mol y pling th si inxing units, i.. tms t i, y pis of th fom (t i, i ) wh h tm t i is qulifi y th ontxt in whih it pps. Th ontxt of ppn of tm is th pth fo nvigting th hihil stutu of th XML oumnt, fom th oot no to th no in whih tm t i ous (.g. in Figu 5, th fist oun of tm John will ssoit with th pth /Amy/Fulty/Dptmnt/Pofsso s its ontxt). Th uthos popos to omput wights of th fom w D (t i, i ) n to xtn quy/oumnt vtos oingly. In ition, thy suggst lxing th quy/oumnt osin simility msu (mo pisly th sl pout pt) y ounting not only fo xt tm in ontxt mthing ut lso fo ontxt smln. Thy mk us of ynmi pogmming LCS (Longst Common Susqun) lgoithm [41] to omput simility vlus twn ontxts (pths), whih susquntly intgt in th osin msu. Stutul tm inxing: Anoth ppoh, xtning th vto sp mol to inopot XML oumnt stutu, is povi y Shli n Muss in [75]. Th uthos xtn th stn notion of tm t i to stutul tm T i, stutul tm ing ll t. Not tht in [75], quis psnt s ll ts (thus inluing stutul tms) n th quy mol is s on t mthing (th uno t inlusion vint is opt [57]) s simpl mns fo fomulting quis without knowing th xt stutu of th XML t. Susquntly, th uthos pt th notions of tm fquny (TF) n invs oumnt fquny (IDF) to th stutul tms: stutul tm T ous in n XML oumnt D if it mths D following th uno t inlusion mthing opto (.g., stutul tm Pofsso/John ous two tims in th XML oumnt in Figu 5. Its TF so w..t. to th oumnt t hn is qul to 2. IDF sos omput in th sm mnn).

10 10 Sumitt to Elsvi Sin Flxil tivl: In ition, s in [36], Shli n Muss nl flxil quying in [75] y intouing th notion of logil XML oumnt. A logil XML oumnt in oumnt olltion is non oth thn su-t of tht olltion (.g. in Figu 5, not only th su-t oot t th Amy no is logil oumnt, ut lso ll its su-ts, tht is th su-ts oot t Fulty, Dptmnt n Lotoy, s wll s h of th Pofsso n Stunt nos). As sult, th us fins, vi h quy, th kin of logil oumnts to tiv. Ths th oumnts hving oots intil to tht of th quy. Similly to [36], TF-IDF sos ll omput t quy tim. Th uthos in [75] lso monstt tht, y justing pmts of th tivl poss, thi mtho n mol oth th lssil vto sp mol (i.. us of lssil tms t i n TF-IDF sttistis) n th oiginl t mthing ppoh (uno t inlusion). Mtix mol: In [62], Pokony n Rjlk onsi tht pvious XML IR mthos, suh s [13], [31], [36], o not suffiintly onsi th stutu of XML oumnts in th tivl poss. In thi stuy, Pokony n Rjlk psnt XML oumnts s mtis inst of simpl vtos. In th vto sp mol, wight w D (t i ) is xpss y l num (.g. TF-IDF so) spifying th istiution of tm t i fo th nti oumnt. In th mtix mol [62], th wight of tm is xpss y vto w D (t i ) 1 k. Suh wight shoul flt th istiution of th tm t i in th XML stutu of th oumnt olltion, w..t. to h pth k ouing in th olltion. Not tht fo simpl pth psnttions, th uthos ly on known DtGui thniqu [34], sin th min i of DtGui is to povi summy of th stutu of oumnt olltion. As oumnts n quis psnt s mtixs, th simility twn quy n oumnt is vlut s th oltion twn osponing mtixs, poviing th sis fo n nhn stutu-w quy systm [62]. Du to its novlty n lvn, XML IR is still vy muh in flux, nw ppohs ing popos gully, whih mks it oviously iffiult to omp th vious mthos. Th itions n to onut th ompison, s wll s th osponing xpimntl fmwok, ontinuously t, nmly in th INEX 1 mpigns it to XML tivl. Consquntly, whil it might sm ly to suvy XML IR t this point in tim, w fl tht simpl psnttion of th popos stuis n thi pplitions woul motivt futh innovtions in th fil. Tl 2 ovs th ntl IR-s XML simility mthos vlop in th littu. 1 INititiv fo th Evlution of XML Rtivl, Oth thniqus fo XML simility Whil ED n IR-s XML ompisons ov wi y of stuis in th littu, svl oth ppohs fo vluting XML simility hv n vlop. Eh of thos xploits iffnt kin of thniqu (.g., tg simility, g mthing, pth simility ) n is it to spifi pplition, som poviing ppoximtions of mo omplx xisting ppohs (minly ED-s). Not tht u to th ivsity of thi unlying thniqus, it is not tivil tsk to tgoiz th vious mthos psnt in this stion. A lssifition s on pplition is oth iffiult n stitiv sin th sm mtho oul xploit in vious pplition omins (spit ing vlop o tst in spifi ontxt). Thus, fo lity of psnttion, w tgoiz mthos following th ntu of th XML t thy tt: stutu-only (isging XML lmnt/ttiut vlus) n stutu-n-ontnt simility mthos Stutu-only XML simility mthos XML lmnt/ttiut vlus gnlly isg whn vluting th stutul poptis of htognous XML oumnts, i.., oumnts oiginting fom iffnt t sous n thus not onfoming to th sm gmm (DTD n/o XML Shm). Suh mthos gnlly suitl fo stutul lssifition/lusting n XML stutul quying pplitions (f. Stion 2.2). Tg simility: Sin optiml ED lgoithms usully qui O(N 2 ) [10] (omplxity with ly lgoithms hing O(N 4 ), s shown in Stion 3.1.2), vious ltntivs n ppoximtions of th ED omputtionl thniqus hv n vlop in th littu, so s to u omplxity. In ptiul, tg simility is onsi s th simplst msu fo XML simility, s it only vluts how losly th st of lmnt/ttiut tgs mth twn two XML oumnts. It ws popos s n ltntiv to mo omplx stutul simility mthos, ptiully ED-s, in th ontxt of XML oumnt lusting [10], [61]. In shot, it onsis th intstion of th sts of tgs, twn th oumnts ing omp, ov th union. Nonthlss, using tg simility, th stutu of th oumnts is ompltly igno, thus ttining low lusting qulity (i.., gnt lusts o not ospon to th pfin ons) in ompison with ED-s mthos [10], [61]. Eg mthing: In [45], Kigl n Shönu omin th simpl no (tg) mthing thniqu (stimting simility twn two XML oumnts s on thi mthing nos, w..t. to spifi no mthing itions silly tg qulity, ompltly ignoing th stutu of th oumnts) n th ED onpt. Th uthos put fow th g mthing ppoh: mthing

11 Sumitt to Elsvi Sin 11 th gs onnting XML nos, thus tking into ount th XML oumnt s stutu in th ompison poss. Th uthos in [45] monstt tht th g mthing ppoh is low oun of th ED thniqus (lss ut), n tht it is of O(E 3 ) omplxity (E is th mximum num of gs in th oumnts ing omp, hving E=N-1 fo XML ts, wh N is th mximum num of nos). A simil g mthing ppoh is povi in [49] wh uthos psnt XML oumnts s it gphs (i. gphs with it gs) n fin istn mti tht ptus th num of ommon gs twn th gph psnttions of two XML oumnts. 1 Egs( G ) Egs( G ) 1 2 Dist( G, G ) = 1 2 (3) Mx{ Egs( G ), Egs( G )} 1 2 Th uthos show tht th popos mti is ffiint in lusting XML oumnts, with spt to ED-s mthos. Th uthos monstt tht thi ppoh is of O(N 2 ) omplxity in th wost s snio n stt tht it usully oms own to O(k N) wh k is smll multipl of N (ll tht N is th mximum num of nos in th XML oumnts ing omp) Pth simility: Th uthos in [10], [64] si th stutu of n XML oumnt s st of pths (stting fom th oot no n ning in th lf nos of th XML oumnt t, tking into ount ll th pths in twn,.g. th pth st of th son XML t in Figu 3.h is {//, //, /, /, /}). Susquntly, XML oumnts omp w..t. thi osponing sts of pths: th mo pths two XML oumnts sh in ommon, th mo simil thy. Th pth simility mtho is shown to of lin tim omplxity [64]. Its omplxity n u to O(1) whn oupl with th shingl thniqu [9] to t onstnt siz psnttions of ity oumnts [10]. XML oumnt lusting xpimnts in [10] n [64] show tht th pth simility mtho povis fily ut simility sults w..t. t ED ompisons. In [42], th uthos xtn vint of th st of pths thniqu n onsi sts of XPths 1. Clssi pths unlin pnt/hil ltionships in th XML oumnt t, ignoing siling infomtion. Nonthlss, XPths (.g., [1]/[1]/[1], [1]/[1]/[1], siing XML oumnt t in Figu 3.h) inopot som siling infomtion. An XPth unlins, fo givn no, how mny ping silings hv th sm ll (It os not ptu siling infomtion out nos whos lls iffnt fom th givn no, whih is xpliitly stt y th uthos). Expimntl sults in [42] show tht th st of XPths ppoh (siing XML oumnts s sts of XPths n onsquntly omping th osponing sts) yils tt lusting qulity thn th lssi st of pths vint onsi in th stuy. 1 St simility: In [11], Cnilli t l. psnt XML oumnts s sts of ttiut-vlus, inluing: th st of pnt-hil ltions (i.., gs), th st of nxt-siling ltions n th st of istint pths stting fom th oot. Using this psnttion, th uthos pfom XML oumnts lssifition n lusting, pplying mthos vlop in [63] n [12] sptivly. Howv, th ppoh is not omp to xisting mthos whih oul lso utiliz fo lssifition/lusting puposs (.g., ED-s, tg simility, g mthing o pth simility). Fst Foui Tnsfom: An oiginl XML simility ppoh, vlop in th ontxt of XML oumnt lusting, is psnt in [29]. H, Fls t l. psnt th stutu of n XML oumnt s tim sis (isging OLTs), h tg oun of n XML lmnt/ttiut osponing to n impuls. Susquntly, thy tmin th g of stutul simility twn oumnts y nlyzing th fqunis of th Fst Foui Tnsfom of osponing tim sis. Th ovll omplxity of Fls t l. s ppoh [29] simplifis to tht of th FFT: O(N log(n)). Howv, th utho in [10] povis n xpimntl itiqu of Fls t l. s FFT mtho [29]. Whil it uns fst thn vint of t ED-s ppohs, Buttl in [10] onlus tht th FFT ppoh os not off n ut msu of simility. Clusting xpimnts, onut on oth l n synthti XML t, show tht th FFT mtho lwys yils th highst o ts (lgst num of mislust oumnts, i.. oumnts put in wong lusts). Stutul simility vi Entopy: In [40], Hlm intous mtho fo msuing th stutul simility twn XML oumnts using ntopy (i.., infomtion istn). Th mtho onsists of two min stps. Fist, th utho xtts stutul infomtion (tg squns, gs n pths) fom th oumnts t hn. Son, th stutul infomtions onning h oumnt X n Y ontnt n thn ompss (otining C(X). C(Y) n C(XY) s th lngths of th ompss fils osponing to th stutul infomtions of oumnts X, Y n thi ontntion sptivly). Th ompssions hn xploit in omputing ntopy: Sim Entopy C( XY ) Min{ C( X ), C( Y )} ( X, Y ) = (4) Mx{ C( X ), C( Y )} Th tionl hin this oiginl mtho is th following: th mo ovlp twn oumnts, th tt th ompssion t will, n thus th high th simility. Th ppoh is of O(N+M) omplxity, wh N n M th sptiv nums of lmnts in th oumnts ing omp. Th utho omp his mtho to on of th min ED-s XML simility ppohs [61], to th Fst Foui Tnsfom mtho [29] s wll s to pth simility vint [10]. Clusting xpimnts show tht th popos mtho pous high oumnt lusting qulity thn [29] n [10], n tht it s on p (n in tin tst ss tt) thn th ED-s mtho in [61].

12 12 Sumitt to Elsvi Sin Stutul pttn mthing: Altntivly to oumnt/oumnt ompison mthos, n ppoh fo oumnt/pttn ompison is povi y Snz t l. in [72]. It is it to stutul nk XML quying, shing fo givn XML pttn (XML t) in oumnt olltion. It stts y mthing n XML pttn to su-ts in th XML t t, tking into ount no ll simility. No lls l 1 n l 2 onsi simil if: l 1 is intil to l 2, l 1 is synonym of l 2 w..t. givn thsuus, o l 1 is synttilly simil to l 2 w..t. to sting mthing thniqu (.g., sting it istn [48]). In susqunt phs, th hihil stutu of nos is onsi to intify, mong th possil mths, thos tht stutully mo simil. Dit inxing stutus (unlining th ll, p-o, post-o n pth of h no) utiliz fo psnting pttns n gions in th oumnt olltion. Timing sults in [72] show tht pfomn is linly pnnt on th siz, in num of nos, of th sult st (w..t. th onsi pttn). XML/DTD simility: A mtho fo msuing th stutul simility twn n XML oumnt n DTD gmm is povi y Btino t l. in [6]. Th popos lgoithm tks into ount th lvl (i.. pth) in whih th lmnts ou in th hihil stutu of th XML n DTD t psnttions. Elmnts t high lvls onsi mo lvnt, in th ompison poss, thn thos t low lvls. Th lgoithm lso onsis lmnt omplxity (i.. th inlity of th su-t oot t th lmnt) whn omputing simility vlus. Th uthos stt tht thi ppoh is of xponntil omplxity. Thy show tht omplxity oms polynomil (O(Γ 2 (N+M)) wh M is th num of nos lmnts/ttiuts in th XML oumnt t, N th num of nos lmnts/ttiuts s wll s?, *, +, An, O optos in th DTD t, n Γ th mximum num of gs out oming fom no of th XML oumnt) if th following ssumption hols: In th ltion of n lmnt, two su-lmnts with th sm tg foin. Th uthos lso povi til isussion of th possil pplitions fo suh n ppoh, minly oumnt lssifition (f. Stion 4) Stutu-n-ontnt XML simility mthos Whil iffnt mthos to XML simility, isging lmnt/ttiut vlus n fousing on th stutul poptis of XML t, hv n popos in th littu, mny oths onsi vlus in thi simility omputtions. Mthos of th ltt goup tgt XML oumnts whih lss stutully ispt (thy might oigint fom th sm t sou, n might vn onfom to th sm gmm), n minly vlop in th ontxts of XML hng mngmnt, t intgtion n stutu-n-ontnt nk quying. Lf no lusting: In th ontxt of XML t intgtion (f. Stion 4), Ling n Yokot povi in [51] n ppoximt XML simility mtho s on lf nos (lf no vlus in ptiul), ntitl LAX (Lflusting s Appoximt XML join lgoithm). Following LAX, th ppoximt simility twn two ts is stimt s th mn vlu of th simility twn thi osponing su-ts (n lgoithm it fo sgmnting XML oumnts into inpnnt su-ts, to tt vi LAX, is lso povi in [51]). Th simility g twn two suts is tmin s th pntg of th num of mth lf nos (pis of lf nos tht hv th sm t vlu) out of th totl num of lf nos in th suts. Th ppoh is of ovll omplxity O(N 2 ) wh N is th mximum num of nos in th XML oumnts ing omp. Expimnts in [51] show tht LAX is, in gnl, fftiv in ssssing XML oumnts simility, w..t. to t ED. Nonthlss, th uthos stt tht whn lg XML oumnts om to ply, i.. whn oumnts hv to fgmnt to fit in min mmoy, simility sults might not optiml (Not tht following LAX, th simility twn two t oumnts pns on thos of thi su-ts. As sult of fgmnttion, su-ts tht sh th lgst similitis i.. mthing su-ts might not tt, h goup of fgmnts ing tt sptly. Hn suoptiml t ompison sults ttin). Rll tht th utho in [18] povis n ppoh (ED-s) pl of omping XML oumnts tht too lg to fit in min mmoy, without ffting th lgoithm s optimlity. Dspit th ft tht th ppohs tgt iffnt pplition omins (th ltt ing pimily vlop fo hng mngmnt puposs), it oul hv n intsting to omp th two mthos. Doumnt List simility: In [43], K n Hus vlop mtho fo omping XML oumnts s oumnts lists. Th ompison poss is untkn in two stps. Fist, h n vy su-t of th oumnt t is tvs, pouing fo h su-t, sting m of th ontnts (vlus) of ll th su-t s lf nos mg togth. Th sult is st of tupls of th fom <pth, ontnt>, on fo h no in th XML t (.g., th tupl osponing to no Dptmnt in Figu 5 is <Amy/Fulty/Dptmnt, John Cm John Tkgi Jms Til >). This psnttion of th XML oumnt is ll oumnt list. Th son stp of th ompison poss onsists in omping th otin oumnt lists, intifying mthing nos (tupls) following thi ontnt n ll similitis (using sting s ompison thniqus,.g., sting it istn [48], s wll s pth similitis (using pth-s ompison mthos suh s th ons si in th Pth simility pgph ov). Nos hving pi-wis simility vlu ov pfin thshol onsis s mthing nos. Not tht uthos only onsi pis of mthing nos to signifint in omputing ovll

13 Sumitt to Elsvi Sin 13 XML oumnt simility (Doumnt simility is omput s th vg of ll simility sos twn mthing nos). Howv, in isussing thi xpimntl sults, thy suggst to lx this ition so s to s oumnt simility w..t. to th num of unmth nos, to gt mo ut ompison sults. Rnking xpimnts onut on (only) synthtilly gnt oumnts. Th uthos xplin tht thy oul not fin l XML t suitl fo thi xpimnts. Ojt Dsiption simility: Wis n Numnn in [86] put fow mtho ntitl Dogmtix fo omping XML lmnts (n onsquntly oumnts) s on thi it vlus, s wll s osponing pnt n hiln similitis. It is vlop in th ontxt of uplit lmnt ttion, i.., intifying lmnts tht psnt th sm l wol ntity, n is g tow t intgtion. Th mtho onsists of th min stps: i) nit ttion, ii) ojt siption n ii) simility omputtion. Th fist stp intifis, in th XML oumnts ing omp, whih lmnts lvnt fo ompison, i.., lmnts tht might si th sm l wol ntity. Hn, pfin mpping twn th lmnts of gmms siing th XML oumnts t hn, n l wol ntitis, is povi s input to th poss. Th son stp oms own to fining th siptions of thos lmnts, ntitl ojt siptions (ODs). Fo givn lmnt 0, th ojt siption ompiss of st of <nm, vlu> tupls unlining 0 s vlu, siling, hiln n/o pnt t, n intifi using it huistis n onitions. Huistis suh s -istnt snnts, onsiing th fist lmnts whih pths in th oumnt o not iff mo thn ius fom lmnt 0 s pth itslf. Fo xmpl, th OD of lmnt Dptmnt (f. Figu 5) following huisti -istn snnts t ius =1 is {<Pofsso, John Cm>, <Stunt, John Tkgi>, <Stunt, Jms Tilo>}. Simil huistis popos fo intifying siling n pnt t siptions. Th thi stp of th ompison fmwok onsists in omping XML lmnts s on thi ojt siptions. Txtul vlus omp using vition of sting it istn [48]. Ovll lmnt similitis vlut using vition of th IDF so (Invs Doumnt Fquny) [69], onsiing th num of mth OD tupls (whih simility is ov givn thshol) ov th totl num of tupls in th two ODs osponing to th lmnts ing omp. Th simility twn two XML oumnts is vlut s tht of thi oot lmnts. Expimntl sults in [86] show tht Dogmtix is fftiv in intifying l n synthti uplit XML lmnts/oumnts. Bysin Ntwoks: Anoth intsting ppoh to uplit ttion is vlop in [47]. It onsis th omplt su-stutu (hiln n snnts) of h lmnt in th oumnts t hn (not only th lmnt s hiln s in [86]). It follows poilisti ppoh, using Bysin ntwok to omin th poilitis of hiln n snnts ing uplits, fo givn pi of XML lmnts in th oumnts ing omp. Th simility twn two XML oumnts ospons to th poilitis of thi oot nos ing uplits. Doumnts ing omp shoul onfom to th sm gmm so s to onstut th Bysin ntwok. Th ltt stitly follows th unlying oumnt gmm. If oumnts not wll fom w..t. to th sm gmm, gmm mthing phs shoul p th onstution of th Bysin ntwok. Th ppoh s omplxity oms own to O(N 2 ) in th wost s snio, wh N is th mximum num of nos in th oumnts ing omp. Expimntl sults in [47] show tht th popos mtho is, in gnl, mo fftiv in tting uplits, in ompison with Dogmtix [86]. Th uthos howv stss on th n fo futh impovmnts, ptiully onning th us of IDF (Invs Doumnt Fquny), whih ws povn fftiv with Dogmtix, ptiully whn th omp oumnts nompss mny lmnts with ummy o pt vlus (Fo instn, th to ol o poution y lmnt vlus, in givn XML oumnt siing movis, not s isiminting in intifying movis, s th movi s titl vlu. This n tkn into ount using IDF, sin th sm ol/y vlus oul pp in mny movis, whs th titl vlu os not). Pttn mthing: An ppoh fo oumnt/pttn ompison, vlop in th ontxt of t intgtion n XML quying, is popos in [26]. Donls t l. mol XML oumnts n pttns s o lll ts, n vlut oumnt/pttn simility using iffnt mtis it to tomi lmnts (i.., lf nos in th XML t) n omplx ons (nompssing oth lmnts, i.., inn nos in th XML t) sptivly. On on hn, uthos onsi tomi lmnt mtis to pnnt on th omins of osponing vlus (txts, ts, nums, ) n thus o not til this issu. On th oth hn, thy istinguish twn omplx olltion n tupl lmnt mtis. Following Donls t l. [26], tupl lmnt is m of iffnt su-lmnts (.g., oot no of T C in Figu 4) whs olltion lmnt nompsss ptitions of th sm su-lmnt (.g., no of T B in figu 4). Mtis povi fo oth typs of omplx lmnts. Ths usivly vlut, xploiting th tomi lmnt mtis, so s to quntify th simility twn th XML pttn n oumnt ts t hn. Not tht this mtho is it to omping oumnts n pttns whih fily simil. Th uthos stt tht th stisftoy vlution of XML lmnt similitis, following thi mtho, quis th omp lmnts to sh th sm ontxts (thi oot pths shoul intil) n hv simil hiln. Authos in [26] o not omp thi mtho s qulity lvls to xisting ppohs, minly thos tgting nk XML quying (.g., IR-s). Tht is poly u to th omplx ntu of suh tsk, s isuss in Stion

14 14 Sumitt to Elsvi Sin Tl 3 pits th vious XML simility mthos isuss ov, long with thi si ftus n pplition omins. 4. Applitions of XML simility Th us of XML simility ngs ov wi sptum of pplitions whih n lssifi in fou mjo goups: i) t whousing, ii) XML lssifition/lusting, iii) XML t intgtion, n iv) nk XML quying Dt whousing: vsion ontol n hng mngmnt On of th min pplitions of XML ompison is to povi suppot fo th ontol of hngs in whous of XML oumnts. In suh ontxt, n ED-s msu (vi n it istn lgoithm, whih n lso intifi in this ontxt s iffning lgoithm, o simply iff), i.. simility msu tht povis lts (silly it sipts), is qui. Th lts off mns to hng ttion n psnttion twn XML oumnts, n onstitut th uiling loks fo XML vsioning [16][17][18][20]. Dlts sml titionl logs in ts systms, n similly to tss, on n fin mny pplitions tht qui ss to logs [55]. Vsions n quying th pst [18][20][55][84]: On my wnt to viw o ss vsion of ptiul oumnt, (pt of) W sit, o th sults of ontinuous quy. This is th stn us of vsions, nmly oing histoy (i.. otin n ol vsion of n XML oumnt). Lt, on might wnt to sk quy out th pst (.g. sk fo th vlu of som XML lmnt t som pvious tim) n to quy hngs (.g. sk fo th list of itms ntly intou in oumnt/tlog). Sin th lts n sto s XML oumnts, suh quis om gul quis ov oumnts. Lning out hngs: Th it istn lgoithm onstuts possil siption of th hngs. It llows to fin, mk-up, n ows hngs twn two o mo vsions of oumnt n lso to upt th ol vsions of th oumnt. This is in th spiit, fo instn, of th Infomtion n Contnt Exhng, ICE 1 [85]. Also, iffnt uss my moify th sm XML oumnt offlin, n lt wnt to synhoniz thi sptiv vsions. Th it istn lgoithm oul us to tt n si th moifitions in o to tt onflits n solv som of thm [20]. Monitoing hngs: In th Xylm pojt fo instn [20], [55], [60], monitoing hngs svs s th fist ft to quy susiption n notifition systms. 1 In th ontxt of ltoni omm, th ICE is stn tht suppots xhnging infomtion out hngs of st of w pgs. It is lso s on lts n snpshots of th t [85]. Th uthos implmnt susiption systm [60] tht llows tting hngs of intst in XML oumnts,.g. tht nw pout hs n to tlogu. At th tim th lt is omput (.g. th it istn lgoithm is xut), th systm vifis fo h tomi hng whth this hng is monito y som susiption (.g. th instion of nw itm in th XML oumnt, ltion of givn itm, t.). Not tht this is lvnt to ICE whih lso povis potool fo notifition. Ahiving: Ahiving is stightfow in this ontxt. It suffis to sto th squn of lts fo tin t to hiv osponing XML t [55]. Mioing: XML ompison, vi it istn lgoithms, n lso us to u th mount of t tnsmitt ov ntwok in mioing pplitions [18]. Popul W n FTP svs oftn hv ozns of mio sits oun th wol. Chngs m to th mst sv n to popgt to th mio sits. Illy, th uss o pogms mking hngs woul kp o of xtly wht t ws upt. Howv, in pti, u to th utonomous n loosly ogniz ntu of suh sits, th is no lil o of hngs [18]. Thfo, ffiint mioing quis iff lgoithms tht omput n popgt only th iffn twn th t vsion t th sv n tht t mio sit XML lssifition n lusting Among th min uss of XML simility/ompison th lssifition n lusting of XML oumnts. XML lssifition: XML simility/ompison nls th lssifition of XML oumnts gth fom th w ginst st of XMl gmms (DTDs o XML shms) l in n XML ts. Th snio povi y Btino t l. [6] ompiss num of htognous XML tss tht xhng oumnts mong h oth, h ts stoing n inxing th lol oumnts oing to st of pfin DTDs. Consquntly, XML oumnts intou in givn ts mth, vi n XML stutul simility mtho, ginst th lol DTDs. Not tht mthing, in suh n pplition, n untkn using n XML oumnt/dtd ompison mtho (lik th on popos in [6] fo msuing th simility twn n XML oumnt n DTD finition) o vi n XML oumnt/oumnt ompison mtho (.g., on of th mthos si pviously fo omping two XML oumnts). Following th ltt sttgy, th DTD will xploit s gnto of XML oumnt stutus (st of possil oumnt stutus vli fo th DTD is onsi). Thn, fo h oumnt stutu, lgoithms fo msuing th stutul simility twn XML oumnts,.g. [18], [23], [61], n ppli. Th mthing sulting in th highst simility vlu n

15 Sumitt to Elsvi Sin 15 onsi s th st nit, th osponing simility vlu ing onsi s th stutul simility so twn th oumnt n th DTD). In suh n pplition, simility thshol shoul intifi (y th us o th systm), unlining th miniml g of simility qui to in n XML oumnt to DTD [6]. Th DTD, fo whih th simility g is highst, n ov th spifi thshol, is slt. Thus, th XML oumnt t hn is pt s vli fo tht DTD. Whn th simility g is low th thshol, fo ll DTDs in th XML ts, th XML oumnt is onsi unlssifi n is sto in positoy of unlssifi oumnts. As sult, non of th pottion, inxing n tivl filitis spifi t DTD lvl n ppli to suh oumnts [6]. XML lusting: Gouping simil XML oumnts togth n impov t stog inxing [76], n thus positivly fft th tivl poss. Fo instn, if two oumnts/lmnts simil, it is likly tht thy oth ith stisfy o not givn quy. Thfo, whn goup togth, simil oumnts/lmnts woul muh si to tiv thn whn stt t iffnt lotions in th stog vi [49]. XML lusting n lso ply mjo ol in fftiv DTD xttion [33]. A lot of XML oumnts foun on th w htognous n lk pfin gmms (DTDs o shms). Nonthlss, hving knowlg of th gmm, fo st of XML oumnts, n vlul fo th pottion, inxing, quying n tivl of ths oumnts [6], [61]. Just s shms nssy in titionl DBMS, th sm is tu fo DTDs n XML tss. Givn olltion of htognous XML oumnts, onstuting singl DTD fo ll ths oumnts woul l to f too gnl finition, whih woul not of muh of us. Howv, whn stutully simil oumnts lust togth fo th DTD xttion poss, mo ut n spifi DTD will onstut fo oumnts in h lust [61]. Clusting n lso itil in infomtion xttion. Cunt infomtion xttion mthos ith impliitly o xpliitly pn on th stutul ftus of oumnts [10]. Bs on stutul lusting, it woul muh si to utomtilly intify th sts of XML oumnts tht usful to infomtion xttion lgoithms, n tht woul pou mningful sults Dt intgtion XML simility/ompison is lso ntl issu in t intgtion. On of th min ftus of XML is tht it n psnt iffnt kins of t fom iffnt t sous, minly on th w. Nonthlss, XML oumnts fom iffnt t sous might ontin nly o xtly th sm infomtion ut might onstut using iffnt stutus. In ition, vn if two oumnts xpss simil ontnts, h of thm my hv som xt infomtion w..t. to th oth. Thus, on ns n fftiv XML simility msu in o to intgt suh t sous, so tht th us n onvnintly ss n qui mo omplt infomtion [38], [50], [51]. Mo pisly, th polm of intgting two XML t sous, fom simility/ompison point of viw, oms own to pfoming n ppoximt join twn ths sous using pfin XML simility msu (most likly t it istn s msu). Givn two XML sous, S 1 n S 2, simility thshol s, n funtion Sim( 1, 2 ) tht sssss th simility twn two oumnts 1 S 1 n 2 S 2, th ppoximt join twn t sous S 1 n S 2 pots in th output ll pis of oumnts ( 1, 2 ) S 1 S 2 suh tht Sim( 1, 2 ) s [37][38]. Susquntly, intgting th intifi pis of oumnts to fom unifi viws of th t n untkn Rnk XML quying With th insing us of XML, spifilly on th w, ffiint tivl of XML oumnts oms mo n mo impotnt [31]. Th ts ommunity hs popos svl lngugs fo quying XML, inluing XML-QL [25], XQL [67] n XQuy [15]. Howv, ths lngugs s on xt mthing n o not suppot nk quis vi txtul/stutul simility. Thfo, svl ttmpts hv n m to xtn ths quy lngugs in o to suppot nk sults, whih is wh XML simility thniqus om to ply. Whil most ppohs in this fmwok s on xtnsions of th vto sp mol, th quy mol us vis with h ppoh. In [31], th uthos xtn XQL, intouing th quy lngug XIRQL whih inopots th notions of tm wights n vgu pits. In [13], Cml t l. voi fining nw XML quy lngug n llow th uss to xpss thi infomtion ns s simpl XML fgmnts (i.. pts of XML oumnts). Th unlying i is to giv lss ontol to th us whn fomulting quis, n to fous most of th logi in th nking mhnism in o to st mt th us s ns (similly to f txt quy mhnisms in titionl IR). In [75], Shli n Muss suppot stutu quis, i.. quis ll ts, n thus giv futh ttntion to XML stutu in th tivl poss. Anoth stuy y Shli [74] intous simpl quy lngug ntitl ppoxql tht suppots hihil Booln-onnt quy pttns. Th intpttion of ppoxql quis is foun on osts quy tnsfomtions wh quis/oumnts mol s ll ts. Thus, t it istn (ppoximt t mthing) vition is us to omput th ost of squn of tnsfomtions twn quy n th t n is us to nk th sults. Appohs ompl to [74] povi in [4], [56]. In ths mo nt woks, th uthos mk us of stutul lxtion on XPth quis, fining spifi lxtion options (g gnliztion i.. tnsfoming pnt/hil g to n nsto/snnt on, lf no ltion ) n

16 16 Sumitt to Elsvi Sin it soing funtions to nl nk quy nswing. An ppoximt nsw to th oiginl XPth us quy is non oth thn n xt mth to on of its lx quis, its so ounting fo th osponing lxtion poss (i.. th so sulting fom pplying th lxtion options to otin th lx quy t hn). Fo instn, th oumnt t in Figu 5 is n ppoximt nsw to th XPth quy //Amy[./Fulty[./Stion]] with so osponing to lting lf no Stion so s to otin th lx quy // Amy [/Fulty]. Vious soing n nking shms popos. In [19], Chinnyng n Kushmik ty to pt xisting nking pilitis in ltionl ts systms to XML tivl. Th uthos put fow ELIXIR, lngug fo XML infomtion tivl tht xtns XML- QL with txtul simility opto. Th osponing quy nswing lgoithm wits th oiginl ELIXIR quis into sis of intmit ltionl t, n mks us of WHIRL 1 [21], [22] to ffiintly vlut th simility opto on this intmit t, susquntly yiling nk XML sults. Similly to [19], Thol n Wikum [81] intou n XML quy lngug, XXL, xtning XML-QL with simility opto. Not tht XXL s simility opto n ppli to lmnt/ttiut nms s wll s to lmnt/ttiut vlus whs ELIXIR [19] is oun to lmnt/ttiut vlus. Th osponing quy posso mks us of th Ol Int-Mi thsuus whil omputing simility. In [8], Bm n Gtz motivt th intoution of nk opto in th XQuy syntx (th sulting quy lngug is intifi s XQuy/IR) nling th us to hoos th simility/ompison mtho to utiliz in th poss (th uthos o not spify th unlying IR thniqu to us). On on hn, XML t whousing, t intgtion n lssifition/lusting pplitions qui ltivly ut XML simility mthos so s to pou tt sults (mo ut hng ttion, mo omplt t intgtion n high qulity lssifition/lusting sptivly). Hn, ppohs in ths pplition s gnlly ED-s (fin-gin). Domin spifi mthos (.g., g mthing, pth simility, som tying to ppoximt mo omplx ED-s mthos) hv n popos (f. Stion 3.3). On th oth hn, most nk XML quying stuis tn to fvou pfomn on uy, iming to pou goo nough sults in sonl tim (inst of tying to gnt pftly ot sults). Thfo, most mthos in this pplition omin IR-s (os-gin). Nvthlss, fw nk XML quying ppohs hv ti to los this gp y xploiting vints of ED [74] n oth thniqus s on pth simility [4], [56]. 1 WHIRL is n infomtion tivl quy lngug it to ltionl t. It inlus txtul simility mti n povis nk simility sults [21], [22]. Thus, pting ED-s mthos (o oth g-s, pth-s,, thniqus) to sh n tiv XML t o, on th oth hn, pting IR-s mthos to t-whousing, t intgtion, lssifition n lusting pplitions oul yil intsting sults (Not tht th i of utilizing ED omputtions in nk quying systm, fo instn, is not novl. It ws intou y Shsh n Zhng in [77] in th ontxt of gni t stutus quying). 5. Disussions n futu sh itions Whil sustntil wok hs n onut oun th XML simility polm, vious issus ging th ffiiny, pfomn n potntil pplitions of XML ompison ppohs yt to tkl. In th min of this stion, w psnt som of ths issus. Fist, w isuss svl limittions of unt ppohs, w..t. th stutul htistis of XML t, whil omping XML oumnts. Aft, w psnt glimps on on of th mgnt polms lt to XML simility: th omintion of stutul n smnti simility ssssmnt whil omping XML t, whih is ing invstigt in oth ED n IR-s ppohs. To onlu this stion, w isuss th usfulnss of XML gmms (DTDs o XML shms) in vloping impov XML ompison mthos XML stutul simility As shown pviously, ng of lgoithms fo omping highly stutu XML oumnts hv n popos in th littu. A thoough invstigtion of th most nt n ffiint XML stutul simility ppohs l us to pinpoint tin ss wh th osponing ompison outom is inut. Usully, XML oumnts n nompss mny optionl n pt lmnts [61]. Suh lmnts inu uing su-ts of simil o intil stutus. As sult, lgoithms fo omping XML oumnt ts shoul w of suh ptitions/smlns so s to ffiintly ssss stutul simility. Not tht in th following, w minly fous ou isussions on ED-s XML ompison lgoithms sin thy tgt igoously stutu t, n thus mo fin-gin w..t. IR-s mthos. Nonthlss, th limittions pinpoint onsquntly tnsitivly ov IR-s XML simility mthos, n oths, s wll Untt Su-t Similitis Ou xmintion of th ppohs povi in [18], [61], [23] l us to intify tin ss wh su-t stutul similitis isg (f. Figu 6): Simility twn ts A/D (su-ts A 1 n D 2 ) in ompison with A/E.

17 Sumitt to Elsvi Sin 17 Simility twn ts F/G (su-ts F 1 n G 2 ) ltivly to F/H. Simility twn ts F/I (su-t F 1 n t I) in ompison with F/J. B 1 G 1 T B F 1 A 1 B 2 h h D 1 D 2 m G 2 f T G T I T D Fig. 6 - Smpl XML ts. Nimn n Jgish in [61] mk us of th ontin in ltion twn ts to ptu su-t similitis, suh s with th A/B n A/C s mntion in Stion (ptition of su-t B 1 ). Nonthlss, whn th ontinmnt ltion is not fulfill, tin stutul similitis igno. Consi, fo instn, ts A n D in Figu 6. Sin D 2 is not ontin in A, it is inst vi fou it options inst of on (inst t), whil tnsfoming A to D, ignoing th ft tht pt of D 2 (sut of nos,, ) is intil to A 1. Thfo, qul istns otin whn omping ts A/D n A/E, isging A/D s stutul smlns (h, w ssum th gnl s wh tomi instion/ltion options of unit osts, =1): Dist(A, D) = Cost Ins (h) + Cost Ins () + Cost Ins () + Cost Ins () + Cost Ins (h) = = 5 Dist(A, E) = Cost Ins (h) + Cost Ins () + Cost Ins (f) + Cost Ins (g) + Cost Ins (h) = = 5 T A T F T C f g C 1 C 2 H 1 m H 2 j T E h f g h E 1 E 2 T H T J Likwis fo th D to A tnsfomtion (t D 2 will not lt vi singl lt t option sin it is not ontin in th stintion t A), hiving Dist(D, A) = Dist(E, A) = 5. Oth typs of su-t stutul similitis tht isg y Nimn n Jgish s ppoh [61] (n likwis miss in [23], [18]) n intifi whn omping ts F/G n F/H, s wll s F/I n F/J. Th F, G, H s is iffnt thn its psso (th A, D, F s) in tht th su-ts shing stutul similitis (F 1 n G 2 ) ou t iffnt pths (whs with A/D, A 1 n D 2 t th sm pth). On th oth hn, th F, I, J s iffs fom th pvious ons sin stutul similitis ou, not only mong su-ts, ut lso t th su-t/t lvl (.g. twn su-t F 1 n t I). Not tht in [23], Dlmgs t l. omplmnt thi it istn lgoithm with ptition/nsting ution poss, summizing th XML oumnts pio to th ompison phs. Suh ution p-possing tnsfoms, fo instn, t B to A (ptition of su-t B 1 ) thus yiling Dist(A, B) = 0 whih is not ut (t A is oviously iffnt thn B). Whil it might usful fo stutul lusting tsks, th ution poss yils inut ompison sults in th gnl s, whih is why it is isg in ou isussion. As fo Dlmgs t l. s ED lgoithm [23], it yils istn vlus intil to th ons tun y Nimn n Jgish s poss [61] in th ov xmpls. Rll tht th lgoithm in [23] is spiliz vsion of tht vlop in [61] wh t instion/ltion osts omput s th sum of th osts of insting/lting ll iniviul nos in th onsi ts Th Spil s of lf no su-ts In ition, non of th ppohs mntion ov is l to fftivly omp oumnts m of pting lf no su-ts. Fo xmpl, following [18][23][61] th sm stutul simility vlu is otin whn omping oumnt K, of Figu 7, to oumnts L n M, Sim(K, L) = Sim(K, M) = 0.5, hving Dist(K, L) = Dist(K, M) = 1. Dist(K, L) = Cost Ins () = 1 Dist(K, M) = Cost Ins () = 1 T K T N T L T O Fig. 7 - XML oumnts onsisting of lf no su-ts. T M T P

18 18 Sumitt to Elsvi Sin Howv, on n liz tht oumnt ts K n L mo simil thn K n M, no of t K pping twi in t L, n only on in M. Likwis fo K/N with spt to K/O n K/P. Intil istns ttin whn omping oumnt ts K/N, K/O n K/P, Dist(K, N)=Dist(K, O)= Dist(K, P)=2, spit th ft tht no is pt th tims in N, twi in O n only pps on in P. Dist(K, N) = Cost Ins () + Cost Ins () = 2 Dist(K, O) = Cost Ins () + Cost Ins () = 2 Dist(K, P) = Cost Ins () + Cost Ins () = 2 sin: W xpliitly mntion th s of lf no ptitions Lf nos spil kin of su-ts: singl no su-ts. Thfo, th issu of su-t smlns n ptitions shoul logilly ov lf nos, so s to ttin mo omplt XML simility ppoh. Lf no ptitions usully s fqunt s sustutu ptitions (i.. non-lf no su-t ptitions) in XML oumnts. Dtting lf no ptitions is spontnous in th XML ontxt, n woul hlp ins th isimintiv pow of XML ompison mthos, s shown in th xmpls of Figu Smnti simility Comining stutul n smnti XML simility is on of th hot topis ntly ing invstigt. Most simility ppohs in th littu fous xlusivly on th stutu of oumnts, ignoing th smntis involv. Howv, in th fil of Infomtion Rtivl, stimting smnti simility twn w pgs is of ky impotn to impoving sh sults [53]. In o to stss th n fo smnti ltnss ssssmnt in XML oumnt ompison, w pot fom [79] th xmpls in Figu 8. <?XML?> <Amy> <Dptmnt> <Lotoy> <Pofsso></Pofsso> <Stunt> </Stunt> </Lotoy> </Dptmnt> </Amy> <?XML?> <Collg> <Dptmnt> <Lotoy> <Ltu></Ltu> </Lotoy> </Dptmnt> </Collg> Fig. 8 - Exmpls of XML oumnts <?XML?> <Ftoy> <Dptmnt> <Lotoy> <Supviso></Supviso> </Lotoy> </Dptmnt> </Ftoy> Using lssil ED omputtions, th sm stutul simility vlu is otin whn oumnt A is omp to oumnts B n C [79]. Howv, spit hving simil stutul htistis, on n oviously ogniz tht smpl oumnt A shs mo smnti htistis with oumnt B thn with C. Fo xmpl, in Figu 8, pis Amy-Collg n Pofsso- Ltu, fom oumnts A n B, smntilly simil whil Amy-Ftoy n Pofsso-Supviso, fom oumnts A n C, smntilly iffnt. Thfo, tking into ount th smnti fto in XML simility omputtions woul oviously mn simility sults. In nt ys, th hv n fw ttmpts to intgt smnti n stutul simility ssssmnt in th XML ompison poss. On of th ly ppohs to popos suh mtho is [81]. Th uthos h mk us of txtul simility opto n utiliz Ol s Int-Mi txt tivl systm to impov XML simility sh. In nt xtnsion of [81] povi in [73], Shnkl t l. popos gni ontologil mol, uilt on WoNt 1, to ount fo smnti simility (inst of utilizing Ol Int-Mi). On th oth hn, th uthos in [6], [72] intify th n to suppot tg simility (synonyms n stms 2 ) inst of tg syntti qulity whil omping XML oumnts. In [79], th uthos stuy th XML smnti simility issu in mo til. Thy onsi th vious smnti ltions nompss in givn fn txonomy/ontology (.g. WoNt) whil omping XML oumnts. Thy intou omin stutul/smnti XML simility ppoh intgting IR smnti simility ssssmnt in titionl ED lgoithm [18]. Nonthlss, th smnti/stutul simility polm is f fom solv. Th vst iffns twn th popos ppohs suggst tht smnti simility oul intgt in multipl wys whil omping XML t. In ition, th smnti omplxity issu, whih is u to ssing th txonomy/ontology onsi, is untly n opn polm. Expimntl sults in [79] onfim th positiv impt of smnti mning on XML simility vlus, whil unlining its hvy impt ging timing omplxity. Thfo, this mgnt topi is likly to thooughly invstigt in th following ys Exploiting XML gmms Anoth possil futu sh ition woul to xplo th us of xisting XML simility mthos to omp, not only th skltons of XML oumnts (lmnt/ttiut lls) ut lso thi infomtion ontnt (lmnt/ttiut vlus). In unt ppohs, whn lmnt/ttiut vlus onsi in th ompison poss [74], [90], thy tt s stings (i.. of t typ Sting) whih is not lwys th s. Vlus oul of Diml, Booln, Sting, Dt typs. Fo h ttyp, iffnt mtho shoul utiliz to omput simility. Thfo, in suh fmwok, XML Shms might hv to intgt in th ompison poss, 1 WoNt is n onlin lxil fn systm (txonomy), vlop t Pinton Univsity NJ USA, wh nouns, vs, jtivs n vs ogniz into synonym sts, h psnting lxil onpt [58] ( 2 Stms signt th mophologil vints of tm: n onym n its xpnsions, singul tm n its plul,

19 Sumitt to Elsvi Sin 19 shms unlining lmnt/ttiut t typs 1 whih qui to omp osponing lmnt/ttiut vlus. A it pplition of suh n ppoh woul th vlopmnt of sophistit XML quy n tivl systm. It woul lso intsting to tkl th XML/Gmm ompison issu, intou in [6]. Sin [6], [80] (to ou knowlg) th only woks tht ov this polm, it sms intsting to look into tht topi. As shown in [6], msuing th stutul simility twn n XML oumnt n DTD hs vious pplitions, inluing XML oumnts lssifition, DTD stutu volution, th vlution of stutul quis, th sltiv issmintion of XML oumnts s wll s th pottion of oumnts. in th pst, xploiting n xtning hivmnts of th omintoil pttn mthing ommunity in t ED lgoithms n lt posss. Howv, thy yt to futh impov n pft s shown in th pvious stion. On th oth hn, IR-s XML simility is ntly gining insing impotn, spilly though th INEX vlution mpigns sponso n ogniz y th IR ommunity. An sin most IR-s mthos mo o lss huisti, thy inssntly isuss, whih psnts n ovwhlming motivtion to vntu in th fil. W liv tht th unifi psnttion of XML simility in this pp will filitt futh sh on th sujt. 6. Conlusion In th pst fw ys, XML hs n stlish s th fto stn fo w pulishing [84], ttting gowing ttntion in ts, infomtion tivl, n mo ntly multimi lt sh (XML is ing insingly us fo siing omplx ojts,.g. multimi infomtion, suh s MPEG-7 2, SVG 3, X3D 4, t.). In this pp, w gv n ovviw of xisting sh lt to XML simility. Th wi ng of ivs mthos popos in th littu w oughly ogniz into th mjo goups: i) ED-s (Eit Distn), ii) IRs (Infomtion Rtivl), n iii) vious oth ontxt n pplition spifi thniqus to XML ompison (som mthos in this goup ppoximtions of mo omplx xisting mthos minly ED-s). Whil IRs mthos tgt XML sh n tivl (spilly fo loosly stutu oumnt-nti XML), ED-s thniqus sm to fous mo on th stutul spt of XML (igoously stutu t-nti viw) n pimily utiliz fo lssifition/lusting n t whousing puposs. W til th possil pplitions of XML ompison posss in vious fils, nging ov t whousing, t intgtion, lssifition/lusting n nk XML quying. In ition, w isuss som possil futu sh itions, oving XML stutul n smnti simility, s wll s th xploittion of XML gmms in vloping impov XML ompison mthos. To onlu, not tht ED-s mthos fo omping XML oumnts hv n thooughly stui 1 XML Shms, lik DTDs, povi mns fo fining th gmm of st of XML oumnts. Howv, shms nl thoough mngmnt of t-typs (19 iffnt t-typs suppot, th us ing l to iv nw t-typs s on th ons tht uiltin), whih is vy stit in DTDs. XML Shm Pt 2: Dttyps Son Eition. 2 Moving Pitus Expts Goup, MPEG WWW Consotium, SVG, 4 W 3D, X3D,

20 Sumitt to Elsvi Sin Eit Distn mthos T. 1 - Chtistis of xisting ED-s XML simility ppohs Appoh XML t tgt 1 Ftus Complxity Applitions Ti [78] Zhng n Shsh [89] Shsh n Zhng [77] Chwth t l. [16] Coén t l. [20] Chwth [18] Nimn n Jgish [61] Dlmgs t l. [23] Tkli t l. [80] Doumnt/Doumnt Stutu-n-ontnt Doumnt/Doumnt Stutu-n-ontnt Doumnt/Doumnt Stutu-only Doumnt/Doumnt Stutu-only Doumnt/Doumnt Stutu-only Doumnt/Gmm (DTD) Stutu-only Fist non-xponntil t ED lgoithm. Inst/lt no nywh in th t, n upt no options. Inst/lt no nywh in th t, n upt no options. Inst/lt no nywh in th t, n upt no options. Inst/lt lf no, upt no, mov no (su-t). Th lgoithm shoul mth spifi itions n ssumptions without whih th sults ttin woul suoptiml. Inst/lt lf no, upt no, mov no (su-t). Som sts of mov options might not optiml. Inst/lt lf no, upt no. Algoithm xtn fo xtnl-mmoy omputtions. Inst/lt lf no, upt no, inst/lt su-t. Outpfoms [Chwth 1999] s lgoithm, whih in tun yils tt stutul lusting sults thn [Zhng n Shsh 1989] s lgoithm. Inst/lt lf no, upt no, inst/lt su-t. Outpfoms [Chwth 1999] s lgoithm. Evluting stutul simility twn XML oumnts n DTD gmms. O( T 1 T 2 pth(t 1) 2 pth(t 2) 2 ) (f. nottions in following pg) O( 2 T 1 min( T 1, T 2 )) O( T 1 T 2 pth(t 1) pth(t 2)) O(n + 2 ) O(N log(n) O(N 2 ) O(N 2 ) O(N 2 ) O(N 3 ) Quying t stutus (ppoximt t y xmpl quis) Vsion ontol n hng mngmnt of smi-stutu t XML Vsion ontol n hng mngmnt Vsion ontol n hng mngmnt of smistutu t XML Stutul lusting XML Stutul lusting XML stutul lssifition T. 2 - Chtistis of xisting IR-s XML simility ppohs Bsi XML Infomtion Rtivl mthos Appoh XML t tgt 2 Ftus Applitions Fuh n Goßjohnn [31] Chinnyng n Kushmik [19] Gs n Shk [36] Cml t l. [13] Shli n Muss [75] Am-Yhi t l. [4] Pokony n Rjlk [62] Doumnt/Quy (i.., Doumnt o pttn) Stutu-n-ontnt Doumnt/Quy Stutu-n-ontnt Doumnt/Quy Stutu-n-ontnt Doumnt/Quy Stutu-n-ontnt Doumnt/Quy Stutu-n-ontnt Doumnt/Quy Stutu-n-ontnt Doumnt/Quy Stutu-n-ontnt Dfining inxing nos. Computing TF-IDF sos lolly. Augmnting wights w..t. th XML stutu. Mks us of xisting king pilitis in ltionl ts systms using WHIRL. Utilizs lssil TF-IDF nking. Buils on [Fuh n Goßjohnn 2001]. Intouing th notions of singl tgoy tivl n multi-tgoy tivl. Flxil tivl: uss spify t quy tim th sop of thi quis. Consis th ontxt of ppn (th oot pth) of tm in omputing TF-IDF sos. Rlxing th osin msu fo omping quy/oumnt vtos y ounting fo ontxt smln. Stutul tms (ll ts) th si inxing units. Flxil tivl: logil XML oumnt XPth quy lxtion. Dfining spifi lxtions options. Soing nsws w..t. quy lxtion poss. Using mtixs inst of vtos to psnt XML oumnts n quis Th istiution of tm is si w..t. its stutul istiution in th XMLoumnt olltion. Rnk XML quying Rnk XML quying Rnk XML quying Rnk XML quying Rnk XML quying Rnk XML quying Rnk XML quying 1 Mthos in this tgoy tgt igoously stutu (t-nti) XML 2 Mthos in this tgoy tgt loosly stutu (oumnt-nti) XML

Learning Schemas for Unordered XML

Learning Schemas for Unordered XML Lning Shms fo Unodd XML Rdu Ciunu Univsity of Lill & INRIA, Fn du.iunu@ini.f S lwk Stwoko Univsity of Lill & INRIA, Fn slwomi.stwoko@ini.f Astt W onsid unodd XML, wh th ltiv od mong silings is ignod, nd

More information

Reading. Minimum Spanning Trees. Outline. A File Sharing Problem. A Kevin Bacon Problem. Spanning Trees. Section 9.6

Reading. Minimum Spanning Trees. Outline. A File Sharing Problem. A Kevin Bacon Problem. Spanning Trees. Section 9.6 Rin Stion 9.6 Minimum Spnnin Trs Outlin Minimum Spnnin Trs Prim s Alorithm Kruskl s Alorithm Extr:Distriut Shortst-Pth Alorithms A Fil Shrin Prolm Sy unh o usrs wnt to istriut il monst thmslvs. Btwn h

More information

Outline. Binary Tree

Outline. Binary Tree Outlin Similrity Srh Th Nikolus Augstn Fr Univrsity of Bozn-Bolzno Fulty of Computr Sin DIS 1 Binry Rprsnttion of Tr Binry Brnhs Lowr Boun for th Eit Distn Unit 10 My 17, 2012 Nikolus Augstn (DIS) Similrity

More information

The pressure screen a tool for the modern fibre alchemist

The pressure screen a tool for the modern fibre alchemist Th pssu sn tool o th mon i lhmist MICHELLE SLOANE 1, PAUL KIBBLEWHITE, MARK RIDDELL 3 AND SUE WILLIAMS 4 1 Sintist Ensis Ppo, Pinipl Sintist Ensis Woo Qulity, 3 Sintist Ensis Woo Qulity, 4 Rsh Oi Ensis

More information

Distributed Systems Principles and Paradigms. Chapter 11: Distributed File Systems. Distributed File Systems. Example: NFS Architecture

Distributed Systems Principles and Paradigms. Chapter 11: Distributed File Systems. Distributed File Systems. Example: NFS Architecture Distriut Systms Prinipls n Prigms Mrtn vn Stn VU mstrm, Dpt. Computr Sin stn@s.vu.nl Chptr 11: Vrsion: Dmr 10, 2012 1 / 14 Gnrl gol Try to mk fil systm trnsprntly vill to rmot lints. 1. Fil mov to lint

More information

XML Data Integration using Fragment Join

XML Data Integration using Fragment Join XML Dt Integtion using Fgment Join Jin Gong, Dvi W. Cheung, Nikos Mmoulis, n Ben Ko Deptment of Compute Siene, The Univesity of Hong Kong Pokfulm, Hong Kong, Chin {jgong,heung,nikos,ko}@s.hku.hk Astt.

More information

Usability Test Checklist

Usability Test Checklist Crtifi Profssionl for Usility n Usr Exprin Usility Tsting (CPUX-UT) Vrsion.0, Jun 0 Pulishr: UXQB. V. Contt: info@uxq.org www.uxq.org Autorn: R. Molih, T. Gis, B. Rumml, O. Klug, K. Polkhn Contnt Lgn...

More information

Hospitals. Internal Revenue Service Information about Schedule H (Form 990) and its instructions is at www.irs.gov/form990.

Hospitals. Internal Revenue Service Information about Schedule H (Form 990) and its instructions is at www.irs.gov/form990. SCHEDULE H Hospitls OMB No. 1545-0047 (Form 990) Complt if th orgniztion nswr "Ys" to Form 990, Prt IV, qustion 20. Atth to Form 990. Opn to Puli Dprtmnt of th Trsury Intrnl Rvnu Srvi Informtion out Shul

More information

Approximate Subtree Identification in Heterogeneous XML Document Collections

Approximate Subtree Identification in Heterogeneous XML Document Collections Approximat Sutr Intiiation in Htrognous XML Doumnt Colltions Ismal Sanz 1, Maro Msiti 2, Giovanna Gurrini 3 an Raal Brlanga 1 1 Univrsitat Jaum I, Spain 2 Univrsità gli Stui i Milano, Italy 3 Univrsità

More information

Instruction: Solving Exponential Equations without Logarithms. This lecture uses a four-step process to solve exponential equations:

Instruction: Solving Exponential Equations without Logarithms. This lecture uses a four-step process to solve exponential equations: 49 Instuction: Solving Eponntil Equtions without Logithms This lctu uss fou-stp pocss to solv ponntil qutions: Isolt th bs. Wit both sids of th qution s ponntil pssions with lik bss. St th ponnts qul to

More information

Schedule C. Notice in terms of Rule 5(10) of the Capital Gains Rules, 1993

Schedule C. Notice in terms of Rule 5(10) of the Capital Gains Rules, 1993 (Rul 5(10)) Shul C Noti in trms o Rul 5(10) o th Cpitl Gins Ruls, 1993 Sttmnt to sumitt y trnsror o shrs whr thr is trnsr o ontrolling intrst Prt 1 - Dtils o Trnsror Nm Arss ROC No (ompnis only) Inom Tx

More information

MANAGEMENT OF INFORMATION SECURITY AND FORENSICS

MANAGEMENT OF INFORMATION SECURITY AND FORENSICS MANAGEMENT OF INFORMATION SECURITY AND FORENSICS CS 307 Ctlog Dsription PREREQUISITE: CS 0. Stuy of informtion surity n igitl fornsis using prtil s stuis. Emphsis is on vloping surity poliis, surity mngmnt

More information

FEE-HELP INFORMATION SHEET FOR DOMESTIC FULL FEE STUDENTS

FEE-HELP INFORMATION SHEET FOR DOMESTIC FULL FEE STUDENTS FEE-HELP INFORMATION SHEET FOR DOMESTIC FULL FEE STUDENTS This is n infomtion sht poducd by th Monsh Lw Studnts Socity Juis Docto Potfolio to ssist full f pying studnts (domstic) in undstnding th issus

More information

fun www.sausalitos.de

fun www.sausalitos.de O ily i f www.lit. Ctt. Cy... 4 5 Rtt... 6 7 B... 8 11 Tt... 12 13 Pt... 14 15. 2 Ctt. Cy. Rtt. B. Tt. Pt Ctt. Cy. Rtt. B. Tt. Pt. 3 Ti t f vyy lif, ity viti. AUALITO i l t t fi, t ty, t t, jy ktil jt

More information

Menu Structure. Section 5. Introduction. General Functions Menu

Menu Structure. Section 5. Introduction. General Functions Menu Menu Structure Section 5 Introduction General Functions Menu Most workstation functions are accessed by menu selections. This section explains the menu structure and provides a tree structured view of

More information

Oracle PL/SQL Programming Advanced

Oracle PL/SQL Programming Advanced Orl PL/SQL Progrmming Avn In orr to lrn whih qustions hv n nswr orrtly: 1. Print ths pgs. 2. Answr th qustions. 3. Sn this ssssmnt with th nswrs vi:. FAX to (212) 967-3498. Or. Mil th nswrs to th following

More information

The Best and Worst of Property Tax Administration

The Best and Worst of Property Tax Administration Th Bst n Wost of Popty Tx Aministtion COST Scoc on Stt Popty Tx Aministtiv Pctics Fick J. Nicly My 2011 Dougls J. Tun Excutiv Summy Fi popty tx ministtion is citiclly impotnt to both iniviul n businss

More information

Operational Procedure: ACNC Data Breach Response Plan

Operational Procedure: ACNC Data Breach Response Plan OP 2015/03 Oprtionl Prour: ACNC Dt Brh Rspons Pln This Oprtionl Prour is issu unr th uthority of th Assistnt Commissionr Gnrl Counsl n shoul r togthr with th ACNC Poliy Frmwork, whih sts out th sop, ontxt

More information

Link-Disjoint Paths for Reliable QoS Routing

Link-Disjoint Paths for Reliable QoS Routing Link-Disjoint Pths or Rlil QoS Routing Yuhun Guo, Frnno Kuiprs n Pit Vn Mighm # Shool o Eltril n Inormtion Enginring, Northrn Jiotong Univrsity, Bijing, 000, P.R. Chin Fulty o Inormtion Thnology n Systms,

More information

11 + Non-verbal Reasoning

11 + Non-verbal Reasoning Prti Tst + Non-vrl Rsoning R th instrutions rfully. Do not gin th tst or opn th ooklt until tol to o so. Work s quikly n s rfully s you n. Cirl th orrt lttr from th options givn to nswr h qustion. You

More information

Uses for Binary Trees -- Binary Search Trees

Uses for Binary Trees -- Binary Search Trees CS122 Algorithms n Dt Struturs MW 11:00 m 12:15 pm, MSEC 101 Instrutor: Xio Qin Ltur 10: Binry Srh Trs n Binry Exprssion Trs Uss or Binry Trs Binry Srh Trs n Us or storing n rtriving inormtion n Insrt,

More information

MATH PLACEMENT REVIEW GUIDE

MATH PLACEMENT REVIEW GUIDE MATH PLACEMENT REVIEW GUIDE This guie is intene s fous for your review efore tking the plement test. The questions presente here my not e on the plement test. Although si skills lultor is provie for your

More information

Enhancing Downlink Performance in Wireless Networks by Simultaneous Multiple Packet Transmission

Enhancing Downlink Performance in Wireless Networks by Simultaneous Multiple Packet Transmission Enhning Downlink Prormn in Wirlss Ntworks y Simultnous Multipl Pkt Trnsmission Zhngho Zhng n Yunyun Yng Dprtmnt o Eltril n Computr Enginring, Stt Univrsity o Nw York, Stony Brook, NY 11794, USA Astrt In

More information

HEAT TRANSFER ANALYSIS OF LNG TRANSFER LINE

HEAT TRANSFER ANALYSIS OF LNG TRANSFER LINE Scintific Jounal of Impact Facto(SJIF): 3.34 Intnational Jounal of Advanc Engining and sach Dvlopmnt Volum,Issu, Fbuay -05 HEAT TANSFE ANALYSIS OF LNG TANSFE LINE J.D. Jani -ISSN(O): 348-4470 p-issn(p):

More information

Load Balancing Algorithm Based on QoS Awareness Applied in Wireless Networks

Load Balancing Algorithm Based on QoS Awareness Applied in Wireless Networks , pp.191-195 http://x.oi.og/10.14257/astl.2015.111.37 Loa Balancing Algoithm Bas on QoS Awanss Appli in Wilss Ntwoks CHEN Xiangqian, MA Shaohui Dpatmnt of Comput Scinc an Tchnology, Hnan Mchanic an Elctical

More information

CompactPCI Connectors acc. to PIGMG 2.0 Rev. 3.0

CompactPCI Connectors acc. to PIGMG 2.0 Rev. 3.0 Ctlog E 074486 08/00 Eition ComptPCI Conntors. to PIGMG.0 Rv. 3.0 Gnrl Lt in 999 PCI Inustril Computr Mnufturrs Group (PICMG) introu th nw rvision 3.0 of th ComptPCI Cor Spifition. Vrsion 3.0 of this spifition

More information

Contracts in outsourcing

Contracts in outsourcing Contt in outouin Chnd, Him M pp, Sptmb 2006 Supvio: D. S. huli Vij Univitit Amtdm Fulty of Sin D olln 1081 1081 HV Amtdm Pf Th M-pp i on of th finl ompuloy ubjt of my tudy uin Mthmti nd nfomti (M t th

More information

Predicting Current User Intent with Contextual Markov Models

Predicting Current User Intent with Contextual Markov Models Priting Currnt Usr Intnt with Contxtul Mrkov Mols Juli Kislv, Hong Thnh Lm, Mykol Phnizkiy Dprtmnt of Computr Sin Einhovn Univrsity of Thnology P.O. Box 513, NL-5600MB, th Nthrlns {t.l.hong, j.kislv, m.phnizkiy}@tu.nl

More information

Level 3. Monday FRACTIONS ⅔ ⅗ 2) ⅔ =?/18. 1) What is a) ⅕ of 30? b) ⅖ of 30?

Level 3. Monday FRACTIONS ⅔ ⅗ 2) ⅔ =?/18. 1) What is a) ⅕ of 30? b) ⅖ of 30? 2014 Th Wkly Pln. All rights rsrv. Mony 2) ⅔ =?/18 1) Wht is ) ⅕ o 30? ) ⅖ o 30? 4) Us or = to show th rltionship twn th ollowing rtions: 3) Writ n quivlnt rtion or ½ ⅔ ⅗ 5) Brook pik ouqut o 24 lowrs.

More information

w ith In fla m m a to r y B o w e l D ise a se. G a s tro in te s tin a l C lin ic, 2-8 -2, K a s h iw a z a, A g e o C ity, S a ita m a 3 6 2 -

w ith In fla m m a to r y B o w e l D ise a se. G a s tro in te s tin a l C lin ic, 2-8 -2, K a s h iw a z a, A g e o C ity, S a ita m a 3 6 2 - E ffic a c y o f S e le c tiv e M y e lo id L in e a g e L e u c o c y te D e p le tio n in P y o d e r m a G a n g re n o su m a n d P so r ia sis A sso c ia te d w ith In fla m m a to r y B o w e l D

More information

1.- L a m e j o r o p c ió n e s c l o na r e l d i s co ( s e e x p li c a r á d es p u é s ).

1.- L a m e j o r o p c ió n e s c l o na r e l d i s co ( s e e x p li c a r á d es p u é s ). PROCEDIMIENTO DE RECUPERACION Y COPIAS DE SEGURIDAD DEL CORTAFUEGOS LINUX P ar a p od e r re c u p e ra r nu e s t r o c o rt a f u e go s an t e un d es a s t r e ( r ot u r a d e l di s c o o d e l a

More information

Algorithmic Aspects of Access Networks Design in B3G/4G Cellular Networks

Algorithmic Aspects of Access Networks Design in B3G/4G Cellular Networks Algorithmi Aspts o Ass Ntworks Dsign in BG/G Cllulr Ntworks Dvi Amzllg, Josph (Si) Nor,DnnyRz Computr Sin Dprtmnt Thnion, Hi 000, Isrl {mzllg,nny}@s.thnion..il Mirosot Rsrh On Mirosot Wy, Rmon, WA 980

More information

Paper Technics Orientation Course in Papermaking 2009:

Paper Technics Orientation Course in Papermaking 2009: P P Otto Cou Pmkg 2009: g to mk u tt you ol o tgt P Wo ould ttd? Otto Cou Pmkg wll b of vlu to t followg gou of ol:- 1. P mll mloy, wo dl dtly wt t o of mkg d w to mov t udtdg of t o d t mll oto t bod

More information

Where preparation meets opportunity. My Academic Planner. Early Academic Outreach Program (EAOP)

Where preparation meets opportunity. My Academic Planner. Early Academic Outreach Program (EAOP) Whr prprtion mts opportunity. My Ami Plnnr Erly Ami Outrh Prorm (EAOP) Follow this 4-stp pln to prpr or mission to th Univrsity o Cliorni (UC), Cliorni Stt Univrsity (CSU) n mny inpnnt olls with similr

More information

ON THE CHINESE CHECKER SPHERE. Mine TURAN, Nihal DONDURMACI ÇİN DAMA KÜRESİ ÜZERİNE

ON THE CHINESE CHECKER SPHERE. Mine TURAN, Nihal DONDURMACI ÇİN DAMA KÜRESİ ÜZERİNE DÜ Fen Bilimlei Enstitüsü Degisi Sı 9 Ağustos 9 On The Chinese Cheke Sphee M. Tun N. Donumı ON THE CHINESE CHECKER SHERE Mine TURAN Nihl DONDURMACI Deptment of Mthemtis Fult of Ats n Sienes Dumlupin Univesit

More information

Physics. Lesson Plan #9 Energy, Work and Simple Machines David V. Fansler Beddingfield High School

Physics. Lesson Plan #9 Energy, Work and Simple Machines David V. Fansler Beddingfield High School Physics Lsson Plan #9 Engy, Wok an Simpl Machins Davi V. Fansl Bingfil High School Engy an Wok Objctivs: Dscib th lationship btwn wok an ngy; Display an ability to calculat wok on by a foc; Intify th foc

More information

Comparing plans is now simple with metal plans. What Does it Mean to Have a 6-Tier Pharmacy Plan? Tie. Individual Health Insurance

Comparing plans is now simple with metal plans. What Does it Mean to Have a 6-Tier Pharmacy Plan? Tie. Individual Health Insurance Compg p ow mp wth mt p Iu p py 0% 0% Iu p py T T T T T T 0% t to M G - No G B No B Iu p py 0% Iu p py Iu Hth Iu Mt p po you wth ho. Th m wth mt p th ptg you p w gy py o. 0% A Hth O tm. Cot. Cg o you. Wh

More information

Revised Conditions (January 2009) LLOYDS BANKING GROUP SHARE ISA CONDITIONS

Revised Conditions (January 2009) LLOYDS BANKING GROUP SHARE ISA CONDITIONS Rvis Conitions (Jnury 2009) LLOYDS BANKING GROUP SHARE ISA CONDITIONS Contnts 1 Who r th prtis?... 2 Wht o wors n phrss in ol typ mn?... 3 Whn i my pln strt?... 4 How o I invst in my pln?... 5 Who owns

More information

Module 5. Three-phase AC Circuits. Version 2 EE IIT, Kharagpur

Module 5. Three-phase AC Circuits. Version 2 EE IIT, Kharagpur Module 5 Three-hse A iruits Version EE IIT, Khrgur esson 8 Three-hse Blned Suly Version EE IIT, Khrgur In the module, ontining six lessons (-7), the study of iruits, onsisting of the liner elements resistne,

More information

C o a t i a n P u b l i c D e b tm a n a g e m e n t a n d C h a l l e n g e s o f M a k e t D e v e l o p m e n t Z a g e bo 8 t h A p i l 2 0 1 1 h t t pdd w w wp i j fp h D p u b l i c2 d e b td S t

More information

Network Decoupling for Secure Communications in Wireless Sensor Networks

Network Decoupling for Secure Communications in Wireless Sensor Networks Ntwork Doupling for Sur Communitions in Wirlss Snsor Ntworks Wnjun Gu, Xiol Bi, Srirm Chllppn n Dong Xun Dprtmnt of Computr Sin n Enginring Th Ohio-Stt Univrsity, Columus, Ohio 43210 1277 Emil: gu, ixi,

More information

1 Fractions from an advanced point of view

1 Fractions from an advanced point of view 1 Frtions from n vne point of view We re going to stuy frtions from the viewpoint of moern lger, or strt lger. Our gol is to evelop eeper unerstning of wht n men. One onsequene of our eeper unerstning

More information

32. The Tangency Problem of Apollonius.

32. The Tangency Problem of Apollonius. . The Tngeny olem of Apollonius. Constut ll iles tngent to thee given iles. This eleted polem ws posed y Apollinius of eg (. 60-70 BC), the getest mthemtiin of ntiquity fte Eulid nd Ahimedes. His mjo wok

More information

SYSTEMS & SERVICES VENDOR PROGRAMS SPECIALTY MARKET PROGRAMS BE A SPECIALIST OR REFER A SPECIALIST

SYSTEMS & SERVICES VENDOR PROGRAMS SPECIALTY MARKET PROGRAMS BE A SPECIALIST OR REFER A SPECIALIST SYSS & SVICS CNUY 21 is th xclusiv sponso in th l stt ctgoy, nd is th only l stt ogniztion tht cn off I ILS wd ils. Poud suppot of st Sls sinc 1979. In 2008, CNUY 21 Cnd ctd th Kids to Cp pog wh vy $2,100

More information

50 MATHCOUNTS LECTURES (10) RATIOS, RATES, AND PROPORTIONS

50 MATHCOUNTS LECTURES (10) RATIOS, RATES, AND PROPORTIONS 0 MATHCOUNTS LECTURES (0) RATIOS, RATES, AND PROPORTIONS BASIC KNOWLEDGE () RATIOS: Rtios re use to ompre two or more numers For n two numers n ( 0), the rtio is written s : = / Emple : If 4 stuents in

More information

1. Number of questions to be answered: ALL Multiple Choice (Section A) and 3 from 5 of the short answer questions (Section B)

1. Number of questions to be answered: ALL Multiple Choice (Section A) and 3 from 5 of the short answer questions (Section B) LEEDS METROPOLITAN UNIVERSITY UK Cntr for Evnts Mngmnt (RESIT) Moul Titl: Evnts Mrkting Ativitis Ami Yr: 2011/12 Lvl: 4 Smstr: 2 Cours: BA(Hons)/ HND Evnt Mngmnt Intrnl Exminrs: Exmintion Dt: 2 n July

More information

Maximum area of polygon

Maximum area of polygon Mimum re of polygon Suppose I give you n stiks. They might e of ifferent lengths, or the sme length, or some the sme s others, et. Now there re lots of polygons you n form with those stiks. Your jo is

More information

December Homework- Week 1

December Homework- Week 1 Dcmbr Hmwrk- Wk 1 Mth Cmmn Cr Stndrds: K.CC.A.1 - Cunt t 100 by ns nd by tns. K.CC.A.2 - Cunt frwrd bginning frm givn numbr within th knwn squnc (instd f hving t bgin t 1). K.CC.B.4.A - Whn cunting bjcts,

More information

N V V L. R a L I. Transformer Equation Notes

N V V L. R a L I. Transformer Equation Notes Tnsfome Eqution otes This file conts moe etile eivtion of the tnsfome equtions thn the notes o the expeiment 3 wite-up. t will help you to unestn wht ssumptions wee neee while eivg the iel tnsfome equtions

More information

Implied volatility formula of European Power Option Pricing

Implied volatility formula of European Power Option Pricing Impli volatility fomula of Euopan Pow Option Picing Jingwi Liu * ing hn chool of Mathmatics an ystm cincs, Bihang Univsity, LMIB of th Ministy of Eucation,, Bijing, 009, P.R hina Abstact:W iv th impli

More information

The Swedish Radiation Protection Institute s Regulations on X- ray Diagnostics;

The Swedish Radiation Protection Institute s Regulations on X- ray Diagnostics; SSI FS 2000:2 Th Swish Rition Prottion Institut s Rgultions on X- ry Dignostis; issu on April 28, 2000. On th sis of 7 of th Rition Prottion Orinn (1988:293) n ftr onsulttion with th Ntionl Bor of Hlth

More information

Campus Sustainability Assessment and Related Literature

Campus Sustainability Assessment and Related Literature Campus Sustainability Assessment and Related Literature An Annotated Bibliography and Resource Guide Andrew Nixon February 2002 Campus Sustainability Assessment Review Project Telephone: (616) 387-5626

More information

Analyzing the Economic Efficiency of ebay-like Online Reputation Reporting Mechanisms Chrysanthos Dellarocas

Analyzing the Economic Efficiency of ebay-like Online Reputation Reporting Mechanisms Chrysanthos Dellarocas Anlyzing th Economic Efficincy of By-lik Onlin Rputtion Rpoting Mchnisms Chysnthos Dllocs Slon School of Mngmnt Msschustts Institut of Tchnology Cmbidg, MA 39, USA dll@mit.du ABSTRACT This pp intoducs

More information

AC Circuits Three-Phase Circuits

AC Circuits Three-Phase Circuits AC Circuits Thr-Phs Circuits Contnts Wht is Thr-Phs Circuit? Blnc Thr-Phs oltgs Blnc Thr-Phs Connction Powr in Blncd Systm Unblncd Thr-Phs Systms Aliction Rsidntil Wiring Sinusoidl voltg sourcs A siml

More information

SEE PAGE 2 FOR BRUSH MOTOR WIRING SEE PAGE 3 FOR MANUFACTURER SPECIFIC BLDC MOTOR WIRING EXAMPLES A

SEE PAGE 2 FOR BRUSH MOTOR WIRING SEE PAGE 3 FOR MANUFACTURER SPECIFIC BLDC MOTOR WIRING EXAMPLES A 0V TO 0V SUPPLY +0V TO +0V RS85 ONVRTR 9 TO OM PORT ON P TO P OM PORT US 9600 U 8IT, NO PRITY, STOP, NO FLOW TRL. OPTO SNSOR # +0V TO +0V RS85 RS85 OPTO SNSOR # PHOTO TRNSISTOR OPTO SNSOR # L TO OTHR Z

More information

SKILL TEST IR(H) HELICOPTER SE ME Application and report form A. Udfyldes af ansøgeren/to be filled out by the applicant:

SKILL TEST IR(H) HELICOPTER SE ME Application and report form A. Udfyldes af ansøgeren/to be filled out by the applicant: SKILL TEST IR(H) HELICOPTER SE ME Applition n rport orm A. Uyls nsørn/to ill out y th pplint: CPR-nr./Dt o Birth: Crtiikt nr./lin no.: (I ny) Ustn Stt/Stt o Lin Issu: Fornvn/First nm(s): Etrnvn/Lst nm:

More information

H ig h L e v e l O v e r v iew. S te p h a n M a rt in. S e n io r S y s te m A rc h i te ct

H ig h L e v e l O v e r v iew. S te p h a n M a rt in. S e n io r S y s te m A rc h i te ct H ig h L e v e l O v e r v iew S te p h a n M a rt in S e n io r S y s te m A rc h i te ct OPEN XCHANGE Architecture Overview A ge nda D es ig n G o als A rc h i te ct u re O ve rv i ew S c a l a b ili

More information

Discovering Petri Nets From Event Logs

Discovering Petri Nets From Event Logs Disovring Ptri Nts From Evnt Logs W.M.P. vn r Alst n B.F. vn Dongn Dprtmnt of Mthmtis n Computr Sin, Thnish Univrsitit Einhovn, Th Nthrlns. {W.M.P.v..Alst,B.F.v.Dongn}@tu.nl Astrt. As informtion systms

More information

WORKING PAPER. Design Of Extended Warranties In Supply Chains

WORKING PAPER. Design Of Extended Warranties In Supply Chains WORKING PAPER No. 09-0 MGT July 009 Dsign Of Etn Waantis In Suly Chains By Kunng Li Sa Houston Stat Univsity Suan Mallik Univsity of Kansas Dili Chhaj Univsity of Illinois at Uana-Chaaign Coyight y Autho

More information

Chapter 7 Yielding criteria

Chapter 7 Yielding criteria Chat 7 Yiling itia. Citia fo yiling What is th maning about yil ition? In this as th stss is un-axial an this oint an aily b tmin. But what if th a sval stss ating at a oint in iffnt ition? Th itia fo

More information

WIRING DIAGRAMS 0. Wire Color. Circuit. Connector. Wire Position in Connector. Wire Identification. Connector Location

WIRING DIAGRAMS 0. Wire Color. Circuit. Connector. Wire Position in Connector. Wire Identification. Connector Location Section Subsection IING IGMS (IING IGMS) IING IGMS The following illustration explains the wires and connectors in wiring diagrams. ire position in connector onnector identification ire color (french)

More information

Forensic Engineering Techniques for VLSI CAD Tools

Forensic Engineering Techniques for VLSI CAD Tools Forensi Engineering Tehniques for VLSI CAD Tools Jennifer L. Wong, Drko Kirovski, Dvi Liu, Miorg Potkonjk UCLA Computer Siene Deprtment University of Cliforni, Los Angeles June 8, 2000 Computtionl Forensi

More information

Combinatorial Testing for Tree-Structured Test Models with Constraints

Combinatorial Testing for Tree-Structured Test Models with Constraints Comintoil Testing fo Tee-Stutued Test Models with Constints Tkshi Kitmu, Akihis Ymd, Goo Htym, Cyille Atho, Eun-Hye Choi, Ngo Thi Bih Do, Yutk Oiw, Shiny Skugi Ntionl Institute of Advned Industil Siene

More information

The art of Paperarchitecture (PA). MANUAL

The art of Paperarchitecture (PA). MANUAL The rt of Pperrhiteture (PA). MANUAL Introution Pperrhiteture (PA) is the rt of reting three-imensionl (3D) ojets out of plin piee of pper or ror. At first, esign is rwn (mnully or printe (using grphil

More information

Volumes by Cylindrical Shells: the Shell Method

Volumes by Cylindrical Shells: the Shell Method olumes Clinril Shells: the Shell Metho Another metho of fin the volumes of solis of revolution is the shell metho. It n usull fin volumes tht re otherwise iffiult to evlute using the Dis / Wsher metho.

More information

c b 5.00 10 5 N/m 2 (0.120 m 3 0.200 m 3 ), = 4.00 10 4 J. W total = W a b + W b c 2.00

c b 5.00 10 5 N/m 2 (0.120 m 3 0.200 m 3 ), = 4.00 10 4 J. W total = W a b + W b c 2.00 Chter 19, exmle rolems: (19.06) A gs undergoes two roesses. First: onstnt volume @ 0.200 m 3, isohori. Pressure inreses from 2.00 10 5 P to 5.00 10 5 P. Seond: Constnt ressure @ 5.00 10 5 P, isori. olume

More information

Last time Interprocedural analysis Dimensions of precision (flow- and context-sensitivity) Flow-Sensitive Pointer Analysis

Last time Interprocedural analysis Dimensions of precision (flow- and context-sensitivity) Flow-Sensitive Pointer Analysis Flow-Insnsitiv Pointr Anlysis Lst tim Intrprocurl nlysis Dimnsions of prcision (flow- n contxt-snsitivity) Flow-Snsitiv Pointr Anlysis Toy Flow-Insnsitiv Pointr Anlysis CIS 570 Lctur 12 Flow-Insnsitiv

More information

G ri d m on i tori n g w i th N A G I O S (*) (*) Work in collaboration with P. Lo Re, G. S av a and G. T ortone WP3-I CHEP 2000, N F N 10.02.2000 M e e t i n g, N a p l e s, 29.1 1.20 0 2 R o b e r 1

More information

Upward Planar Drawings of Series-Parallel Digraphs with Maximum Degree Three

Upward Planar Drawings of Series-Parallel Digraphs with Maximum Degree Three Upwr Plnr Drwins of ris-prlll Dirps wit Mximum Dr Tr (Extn Astrt) M. Aul Hssn m n M. iur Rmn Dprtmnt of Computr in n Eninrin, Bnls Univrsity of Eninrin n Tnoloy (BUET). {sm,siurrmn}@s.ut.. Astrt. An upwr

More information

Industry regulations Jurisdictional regulations Legal defensibility Legal frameworks Legal research

Industry regulations Jurisdictional regulations Legal defensibility Legal frameworks Legal research A Dutis, Tsks, n Stps Mnging Informtion Risk n Complin 1 Monitor lgl n rgultory lnsp Engg with lgl prtmnt n othr stkholrs Intify n intrprt xisting pplil lws of ll jurisitions n rgultions Intify rsours

More information

Standard Conditions for Street Traders The Royal Borough of Kensington and Chelsea. Revised standard conditions for street trading

Standard Conditions for Street Traders The Royal Borough of Kensington and Chelsea. Revised standard conditions for street trading Stnr Conitions or Strt Trrs Th Royl Borough o Knsington n Chls Rvis stnr onitions or strt tring Th Royl Borough o Knsington n Chls strt tring linss stnr onitions 2006 1 Dinitions Th ollowing xprssions

More information

Quick Guide to Lisp Implementation

Quick Guide to Lisp Implementation isp Implementtion Hndout Pge 1 o 10 Quik Guide to isp Implementtion Representtion o si dt strutures isp dt strutures re lled S-epressions. The representtion o n S-epression n e roken into two piees, the

More information

Recall from Last Time: Disjoint Set ADT

Recall from Last Time: Disjoint Set ADT Ltur 21: Unon n Fn twn Up-Trs Toy s An: Plntn n rown orst o Up-Trs Unon-n n Fn-n Extn xmpl Implmntn Unon/Fn Smrt Unon n Fn Unon-y-sz/t n Pt Comprsson Run Tm Anlyss s tou s t ts! Covr n Cptr 8 o t txtook

More information

Victims Compensation Claim Status of All Pending Claims and Claims Decided Within the Last Three Years

Victims Compensation Claim Status of All Pending Claims and Claims Decided Within the Last Three Years Claim#:021914-174 Initials: J.T. Last4SSN: 6996 DOB: 5/3/1970 Crime Date: 4/30/2013 Status: Claim is currently under review. Decision expected within 7 days Claim#:041715-334 Initials: M.S. Last4SSN: 2957

More information

Summary: Vectors. This theorem is used to find any points (or position vectors) on a given line (direction vector). Two ways RT can be applied:

Summary: Vectors. This theorem is used to find any points (or position vectors) on a given line (direction vector). Two ways RT can be applied: Summ: Vectos ) Rtio Theoem (RT) This theoem is used to find n points (o position vectos) on given line (diection vecto). Two ws RT cn e pplied: Cse : If the point lies BETWEEN two known position vectos

More information

Summary of changes to Regulations recommended to the Senate by Graduate School Management Committee. Changed wording is shown in bold italics.

Summary of changes to Regulations recommended to the Senate by Graduate School Management Committee. Changed wording is shown in bold italics. Summry of hngs to Rgultions rommn to th Snt y Grut Shool Mngmnt Committ. Chng woring is shown in ol itlis. Gnrl Rgultions for Dgrs y Rsrh n Thsis 1. Inlusion of th Dotor of Miin (MD) wr throughout. 2.

More information

Scholarship Help for Technology Students

Scholarship Help for Technology Students i NOVEMBER 2014 Sli Hl f Tl S S i il ili l j i il i v f $150000 i li VN l f li Pl Tl N f xl i ii f v Pi Oli i N fi f i f vl i v f f li f i v f Viii Sli f vill f flli j: Pl Tl Mi Alli Hl li A Ifi Tl li

More information

i n g S e c u r it y 3 1B# ; u r w e b a p p li c a tio n s f r o m ha c ke r s w ith t his å ] í d : L : g u id e Scanned by CamScanner

i n g S e c u r it y 3 1B# ; u r w e b a p p li c a tio n s f r o m ha c ke r s w ith t his å ] í d : L : g u id e Scanned by CamScanner í d : r ' " B o m m 1 E x p e r i e n c e L : i i n g S e c u r it y. 1-1B# ; u r w e b a p p li c a tio n s f r o m ha c ke r s w ith t his g u id e å ] - ew i c h P e t e r M u la e n PACKT ' TAÞ$Æo

More information

A n d r e w S P o m e r a n tz, M D

A n d r e w S P o m e r a n tz, M D T e le h e a lth in V A : B r in g in g h e a lth c a r e to th e u n d e r s e r v e d in c lin ic a n d h o m e A n d r e w S P o m e r a n tz, M D N a tio n a l M e n ta l H e a lth D ir e c to r f

More information

Functional Valuation of Ecosystem Services on Bonaire

Functional Valuation of Ecosystem Services on Bonaire Funtionl Vlution of Eosystm Srvis on Bonir - An ologil nlysis of osystm funtions provi y orl rfs - Ingri J.M. vn Bk MS Aquultur n Fishris Spilistion Mrin Rsours n Eology Wgningn Univrsity Th Nthrlns MS-

More information

Dinh Hong Giang 1,2, Ed Sarobol 2, * and Sutkhet Nakasathien 2 ABSTRACT

Dinh Hong Giang 1,2, Ed Sarobol 2, * and Sutkhet Nakasathien 2 ABSTRACT Kstsrt J. (Nt. Si.) 49 : 1-12 (215) Et o Plnt Dnsity n Nitrogn Frtilizr Rt on Growth, Nitrogn Us Eiiny n Grin Yil o Dirnt Miz Hyris unr Rin Conitions in Southrn Vitnm Dinh Hong Ging 1,2, E Srool 2, * n

More information

IncrEase: A Tool for Incremental Planning of Rural Fixed Broadband Wireless Access Networks

IncrEase: A Tool for Incremental Planning of Rural Fixed Broadband Wireless Access Networks InrEs: A Tool or Inrmntl Plnning o Rurl Fix Bron Wirlss Ass Ntworks Giomo Brnri n Mhsh K. Mrin Shool o Inormtis Th Univrsity o Einurgh, UK Frnso Tlmon n Dmitry Rykovnov EOLO L NGI SpA, Miln, Itly Astrt

More information

A122 MARION COUNTY HEALTH BUILDING HVAC, GLAZING AND LIGHTING RENOVATION 75% DOCUMENTS 08/31/2015

A122 MARION COUNTY HEALTH BUILDING HVAC, GLAZING AND LIGHTING RENOVATION 75% DOCUMENTS 08/31/2015 7 ' 7 /" ' " ' /" ' 9 /" ' 0" ' 0" ' 0" ' 0" ' " ' /" 0 NRL SHT NOTS IL VRIY XISTIN PRIOR TO WORK N NOTIY RHITT/NINR O ISRPNIS TWN RWINS N XISTIN ONITIONS. 0 0 0 PTH LOTIONS N IR PROOIN WHR XISTIN WLLS

More information

The invention of line integrals is motivated by solving problems in fluid flow, forces, electricity and magnetism.

The invention of line integrals is motivated by solving problems in fluid flow, forces, electricity and magnetism. Instrutor: Longfei Li Mth 43 Leture Notes 16. Line Integrls The invention of line integrls is motivted by solving problems in fluid flow, fores, eletriity nd mgnetism. Line Integrls of Funtion We n integrte

More information

1. Definition, Basic concepts, Types 2. Addition and Subtraction of Matrices 3. Scalar Multiplication 4. Assignment and answer key 5.

1. Definition, Basic concepts, Types 2. Addition and Subtraction of Matrices 3. Scalar Multiplication 4. Assignment and answer key 5. . Definition, Bsi onepts, Types. Addition nd Sutrtion of Mtries. Slr Multiplition. Assignment nd nswer key. Mtrix Multiplition. Assignment nd nswer key. Determinnt x x (digonl, minors, properties) summry

More information

(Ch. 22.5) 2. What is the magnitude (in pc) of a point charge whose electric field 50 cm away has a magnitude of 2V/m?

(Ch. 22.5) 2. What is the magnitude (in pc) of a point charge whose electric field 50 cm away has a magnitude of 2V/m? Em I Solutions PHY049 Summe 0 (Ch..5). Two smll, positively chged sphees hve combined chge of 50 μc. If ech sphee is epelled fom the othe by n electosttic foce of N when the sphees e.0 m pt, wht is the

More information

Words Symbols Diagram. abcde. a + b + c + d + e

Words Symbols Diagram. abcde. a + b + c + d + e Logi Gtes nd Properties We will e using logil opertions to uild mhines tht n do rithmeti lultions. It s useful to think of these opertions s si omponents tht n e hooked together into omplex networks. To

More information

Performance Control of PMSM Drives Using a Self-tuning PID

Performance Control of PMSM Drives Using a Self-tuning PID fom oo of MSM U Sf- I XIO X I Yoo I M pm of h U j 8 h ho: 86--67858 Fx: 86--67896 -m: xo_x@h.. -I h pp f- I oo o mmm mho pp o h pm-m hoo moo MSM p oo m. h popo mho h ppoxm mmm pop h of h o-oop m po-m whh

More information

Chapter 3 Chemical Equations and Stoichiometry

Chapter 3 Chemical Equations and Stoichiometry Chptr Chmicl Equtions nd Stoichiomtry Homwork (This is VERY importnt chptr) Chptr 27, 29, 1, 9, 5, 7, 9, 55, 57, 65, 71, 75, 77, 81, 87, 91, 95, 99, 101, 111, 117, 121 1 2 Introduction Up until now w hv

More information

Calculating the Real ROI of Video Conferencing Technology. SHOCKING TRUTH It s Not Just About Travel Costs!

Calculating the Real ROI of Video Conferencing Technology. SHOCKING TRUTH It s Not Just About Travel Costs! Clcultig th Rl ROI f Vid Cfcig Tchlgy SHOCKING TRUTH It s Nt Just Abut Tvl Csts! Clcultig th Rl ROI f Vid Cfcig Tchlgy Svig M Th Just th Cst f Tvl...3 Old Schl Th Tditil Wy f Clcultig VC ROI...4 Tim =

More information

CHAPTER 31 CAPACITOR

CHAPTER 31 CAPACITOR . Given tht Numer of eletron HPTER PITOR Net hrge Q.6 9.6 7 The net potentil ifferene L..6 pitne v 7.6 8 F.. r 5 m. m 8.854 5.4 6.95 5 F... Let the rius of the is R re R D mm m 8.85 r r 8.85 4. 5 m.5 m

More information

Inter-domain Routing

Inter-domain Routing COMP 631: COMPUTER NETWORKS Inter-domin Routing Jsleen Kur Fll 2014 1 Internet-sle Routing: Approhes DV nd link-stte protools do not sle to glol Internet How to mke routing slle? Exploit the notion of

More information

SCO TT G LEA SO N D EM O Z G EB R E-

SCO TT G LEA SO N D EM O Z G EB R E- SCO TT G LEA SO N D EM O Z G EB R E- EG Z IA B H ER e d it o r s N ) LICA TIO N S A N D M ETH O D S t DVD N CLUDED C o n t e n Ls Pr e fa c e x v G l o b a l N a v i g a t i o n Sa t e llit e S y s t e

More information

P U B L I C A T I O N I N T E R N E 1800 PARTIAL ORDER TECHNIQUES FOR DISTRIBUTED DISCRETE EVENT SYSTEMS: WHY YOU CAN T AVOID USING THEM

P U B L I C A T I O N I N T E R N E 1800 PARTIAL ORDER TECHNIQUES FOR DISTRIBUTED DISCRETE EVENT SYSTEMS: WHY YOU CAN T AVOID USING THEM I R I P U B L I C A T I O N I N T E R N E 1800 N o S INSTITUT DE RECHERCHE EN INFORMATIQUE ET SYSTÈMES ALÉATOIRES A PARTIAL ORDER TECHNIQUES FOR DISTRIBUTED DISCRETE EVENT SYSTEMS: WHY YOU CAN T AVOID

More information

Data Quality Certification Program Administrator In-Person Session Homework Workbook 2015-2016

Data Quality Certification Program Administrator In-Person Session Homework Workbook 2015-2016 Dt Qulity Certifition Progrm Aministrtor In-Person Session Homework Workook 2015-2016 Plese Note: Version 1.00: Pulishe 9-1-2015 Any exerises tht my e upte fter this printing n e foun online in the DQC

More information

Practice Test 2. a. 12 kn b. 17 kn c. 13 kn d. 5.0 kn e. 49 kn

Practice Test 2. a. 12 kn b. 17 kn c. 13 kn d. 5.0 kn e. 49 kn Prtie Test 2 1. A highwy urve hs rdius of 0.14 km nd is unnked. A r weighing 12 kn goes round the urve t speed of 24 m/s without slipping. Wht is the mgnitude of the horizontl fore of the rod on the r?

More information

University of Mumbai Application for selection as Best College

University of Mumbai Application for selection as Best College Univrsity o Mumi Applition or sltion s Bst Collg 1 Inormtion o th Sponsoring Soity Nm o th Sponsoring Soity with Rgistrtion No. Nms o th Bor o Trusts Whthr Sponsoring Soity is running ny othr Institutions

More information

Attachment 1 Package D1-1 (Five (5) Locations) 9-26-13Revised 11-1-13

Attachment 1 Package D1-1 (Five (5) Locations) 9-26-13Revised 11-1-13 Space Identifier Near Gate ttachment Package - (Five (5) Locations) 9-26-3Revised --3 Proposed oncept Square Footage Minimum nnual Guarantee Term in Years --Z0 ustoms urrency xchange 98 $20,500 75-2-S06

More information

Tank Level GPRS/GSM Wireless Monitoring System Solutions

Tank Level GPRS/GSM Wireless Monitoring System Solutions Tank Lvl GPRS/GSM Wilss Monitoing Systm Solutions HOLYKELL TECHNOLOGY CO.LTD May,2014 Ⅰ. Solution Rquimnts 1. Intoduction Th solution is mainly including: wilss data tansciv tminal, lvl snso and PC sv

More information

Orbits and Kepler s Laws

Orbits and Kepler s Laws Obits nd Keple s Lws This web pge intoduces some of the bsic ides of obitl dynmics. It stts by descibing the bsic foce due to gvity, then consides the ntue nd shpe of obits. The next section consides how

More information