Candidate Generation for Interactive Chinese Speech Recognition

Size: px
Start display at page:

Download "Candidate Generation for Interactive Chinese Speech Recognition"

Transcription

1 Canddat Gnraton for ntractv Chns Spch Rcognton Xnhu L 1 2 Xangdong Wang 1 Yulang Qan 1 Shouxun Ln 1 1 Ky Laboratory of ntllgnt nformaton Procssng nsttut of Computng Tchnology Chns Acadmy of Scncs Bng Chna 2 Graduat Unvrsty of Chns Acadmy of Scncs Bng Chna { lxnhu xdwang ylqan sxln}@ct.ac.cn Abstract Dspt progrss n automatc spch rcognton usr nvolvmnt and assstanc ar stll ndd to achv prformanc smlar to human n ral-world tasks. Though many ntractv spch rcognton systms usng canddat slcton ar dvlopd for Englsh and Japans smlar systms for Chns fac partcular dffcults du to charactrstcs of th Chns languag spcally at th canddat gnraton stag. n ths papr a mthod s proposd for canddat gnraton n ntractv Chns spch rcognton whch lsts all compttv canddats n forms of Chns charactrs n stad of words. Th man da s to algn tm-ovrlappd lnks n Chns word lattc nto clustrs basd on thr phontc smlarty and thn obtan canddats by splttng words attachd to lnks nto Chns charactrs. Exprmntal rsults dmonstratd that th canddats gnratd by th proposd mthod can lad to hgh prformanc n ntractv Chns spch rcognton. Kywords: spch rcognton rror corrcton canddat gnraton ntractv spch rcognton. 1. ntroducton Automatc spch rcognton has mad grat progrss ovr th past fw dcads. Howvr for tasks such as larg vocabulary contnuous spch rcognton (LVCSR) th stat-of-art tchnqu s stll far from ral-world applcaton. As a mattr of fact t sms vry dffcult for computr to achv prformanc smlar to human through th mprovmnt of rcognton tchnology n nar futur. Ths mpls that nvolvmnt and assstanc of human usrs ar ndd to achv hgh accuracy n spch rcognton. W rfr to spch rcognton wth nvolvmnt and assstanc of usrs as ntractv spch rcognton. A maor tchnqu n ntractv spch rcognton s rror corrcton or spch rpar va canddat slcton [1-4]. Undr th schm of canddat slcton th spch rcognton systm outputs not only th bst word squnc as rcognton rsult but also a lst of compttv canddats for ach word (not for a sntnc) and th usr can corrct th rcognton rrors by slctng rght canddats. For xampl a canddat lst for th uttranc w nd som hlp s shown n Fg. 1. Each word n th bst word squnc s sparatd and ts compttv canddats ar lstd blow t. W rfr to ach word and ts compttv canddats as a rcognton sgmnt. Th usr can corrct a rcognton rror by slctng th rght compttv canddat n th sgmnt. t s rportd that much mprovmnt n accuracy can b achvd by canddat slcton [4]. Bst squnc Compttv canddat Sgmnt WE NEED SUN HELP WLL LEAD SOME SON Fgur 1. Canddat lst for uttranc w nd som hlp. For th gnraton of compttv canddats most systms us th confuson ntwork (CN) algorthm [5-7] whch obtans th canddats by transformng word lattc nto CN. Durng th dcodng stag spch rcognton systms gnrat a word lattc as an ntrmdat rsult and sarch n th lattc for a bst word squnc as fnal rcognton rsult. A word lattc s a Drctd Acyclc Graph (DAG) whch contans a larg numbr of word hypothss dnotd by corrspondng lnks and thr assocatd lklhood scors. Fg. 2 gvs an xampl of an Englsh word lattc. To form th confuson ntwork lnks on a word lattc ar algnd accordng to tm nformaton and th lattc s transformd nto a lnar graph. Th transformaton s prformd by a clustrng procdur that groups tm ovrlappd lnks nto clustrs basd on thr phontc smlarty and word probablts whl prsrvng th prcdnc ordr of th lnks ncodd n th word lattc. Fg. 3 gvs th CN obtand by transformng th word lattc of Fg /09/$ EEE 583 1

2 SL VERY FNE T HAVE VEAL FAST SL MOVE VERY SL OFTEN T Fgur 2. An Englsh word lattc. HAVE MOVE T VEAL VERY FNE OFTEN FAST Fgur 3. CN of th word lattc n Fg. 2. Though thr ar many ntractv spch rcognton systms usng canddat slcton for Englsh and Japans [1-4] as to our knowldg thr ar no smlar systms for Chns spch yt. Th man challng for canddat slcton to b usd n Chns spch rcognton ls n that th CN algorthm cannot b usd for canddat gnraton du to charactrstc of th Chns languag. n Chns word lattc th word hypothss attachd to ach lnk may b a sngl Chns charactr or a strng of Chns charactrs (n gnral nclud 1-4 sngl Chns charactrs) and compttv rlatonshp may xst btwn two charactrs whch ar both parts of dffrnt word hypothss. Thrfor t s mpossbl to group all th word hypothss wth compttv rlatonshp nto a clustr by transformng Chns word lattc nto Chns CN. Fg. 4(a) gvs an xampl of Chns word lattc and Fg. 4(b) shows th Chns CN obtand by transformng th word lattc of Fg. 4(a). n Fg. 4(b) th word hypothss compts not only wth othr word hypothss n th sam clustr but also wth othr word hypothss n othr clustrs (lk bold word hypothss n th Fgur). n fact du to th charactrstc of Chns languag that both a sngl charactr and a strng of charactr can form a word t s dsrd by usrs that canddats can b lstd n trms of charactrs n stad of words whch cannot b ralzd by usng CN. n ordr to gnrat sutabl canddats for ntractv Chns spch rcognton consdrng th charactrstc of th Chns languag w propos an approach of canddat gnraton for ntractv Chns spch rcognton whch lsts all compttv canddats n forms of Chns charactrs n stad of words. Frst smlar to gnraton of CN tmovrlappd lnks n Chns word lattc ar algnd nto clustrs basd on thr phontc smlarty. And thn canddats ar obtand by splttng words attachd to lnks nto Chns charactrs. Exprmntal rsults show that th canddats gnratd by th proposd mthod can achv hgh covrag of corrct charactrs rsultng n hgh prformanc n ntractv Chns spch rcognton. Fgur 4. Chns word lattc and corrspondng CN. Th rst of th papr s organzd as follows. Scton 2 ntroducs th man da and procdur of th proposd mthod. n Scton 3 algorthms for canddat gnraton ar dscrbd n dtal. Exprmntal rsults of th proposd mthod ar prsntd n Scton 4. And fnally conclusons ar drawn and futur work ar dscussd n Scton Man da of th Proposd Approach 2.1. Prncpls of Canddat Gnraton n ordr to nabl usrs to quckly and ffcntly corrct rcognton rrors by canddat slcton th canddat should mt th followng prncpls: 1. Canddats wth compttv rlatonshp should b groupd nto th sam rcognton sgmnt. n ths cas th usr can fnd th rght compttv canddat wthn a small ara. 2. Th ordr of th rcognton sgmnts should b consstnt wth rcognton tm. 3. Th compttv canddats of ach rcognton sgmnt should b sortd by th valu of th postror probablts whch can b computd from word lattc. Ths mans th plausbl canddats ar locatd at uppr postons n th rcognton sgmnt. t can b sn that for Chns spch th CN gnratd from Chns word lattc dos not follow prncpls 1 and 3 snc compttv rlatonshp actually xsts btwn charactrs nstad of words n Chns Procdur of th Proposd Mthod n ordr to gnrat Chns canddats that mt th prncpls of canddat gnraton w propos a 584 2

3 Chns canddat gnraton mthod whch lsts canddats n charactrs. Frst an algnd ntwork s gnratd by algnng th lnks n Chns word lattc and thn charactr canddats ar gnratd by splttng Chns words n th algnd ntwork nto charactrs. Fg. 5 gvs an llustraton of Chns canddat gnraton. algnd clustrs n algnd ntwork and k dnots th st of lnks n th k-th clustr. (3) Chns canddats A Chns canddat lst s dnotd by C { C 0 C 1 C 2 C L } whch s th st of all rcognton sgmnts n Chns canddat whr C l { c0 c1 c2 ck} dnots th st of all canddats n th l-th rcognton sgmnt. Each c k C s dfnd as { W P } whr s th l word of canddat s th postror probablty c k of canddat. c k P ck ck ck W ck 3.2. Algorthm for Algnmnt of Chns Word Lattc n th frst stag th word lattc s algnd to form a lnar algn ntwork by clustrng lnks n th lattc. Th lnks whch mt th followng condtons ar clustrd nto on class: (1) th last Chns charactrs attachd to ach lnk ar phontcally smlar; (2) Th lnks ar tm-ovrlappd. Th algnmnt of Chns word lattc can b dscrbd as follows: 3. Algorthms for Canddat Gnraton for ntractv Chns Spch Rcognton 3.1. Assumpton and Dnotaton Frst w gv som assumptons and dnotatons usd by followng dscrpton. (1) Chns word lattc A Chns word lattc s dnotd by L=<N E> whr N { n0 n1 n2 n } s th st of all nods n Chns word lattc and E { J } s th st of all lnks n Chns word lattc. n N tn ( ) dnots th tm mark. Each k E s dfnd as { S F W A L } whr s k k k k k k F k th start nod of lnk s th nd nod of lnk W k s th word on th lnk k and ar th k k acoustc lklhood scor and th languag modl scor rspctvly. (2) Algnd ntwork An algnd ntwork s dnotd by { } whch s th st of all E E E E E A Fgur 5. Ovrvw of Chns canddat gnraton K A L S k k Stp-1: Calculat th postror probablty p() of ach lnk n th lattc usng forwardbackward algorthm [5]. Stp-2: Sort all lnks n E by th ordr of ncrasng tf) ( k and sort th lnks wth sam tf ( k ) by th ordr of ncrasng ts ( k ). Stp-3: ntalz algnd ntwork: for ach lnk E f ts ( ) 0 0. Stp-4: For ach lnk E 1 2 suppos that th prvous lnk s dnotd by 1 (a) f ts ( ) ts ( ) and tf ( ( ) 1 ) tf 1. (b) f E such that ts ( ) tf) ( 1 1. Also for ach lnk f SM( ) SM( ) 1 1 and \ whr SM( ) smc (() c ( ) ) ovrlap ( ) whr c( ) and c( ) ar th last Chns charactrs corrspondng to and sm (..) s th phontc smlarty btwn two charactrs computd from th most lkly phontc bas forms and ovrlap( ) s 585 3

4 dfnd as th tm ovrlap btwn and normalzd by th sum of thr lngths. (c) f K and K such that mn{ u ( El lk1 ts ( ) tf ( ) and )} ) u ( whr u() s th numbr of Chns charactrs attachd to th lnk. (d) f and K such that ts ( ) tf ( ) and mn{ u ( )} u ( ) K 1 1. () f and K such that K ts ( ) tf ( ) E 1 l lk and mn{ u ( )} u ( ) E 1 l lk H H K H whr H s dtrmnd by th followng quaton 1 H arg max{ SM( )} KH we ( ) H E H whr w ( H) s th numbr of lnks n H SM( ) rman th sam as bfor. Stp-5: n ach algnd clustr mrg all lnks wth th sam Chns word nto a sngl lnk whos postror probablty quals to th sum of postror probablts of all lnks mrgd. whr QW ( s th -th charactr n th ) Chns word attachd to lnk. (b) f u( ) > num lt th word of canddat c b W c = QW ( and th postror ) probablty of canddat c b Pc = p ( ) st C m num u ( ) C mnum u( ) c 01 u( ) 1. Stp-3: nn1 m m num. f n w( E A ) rturn to Stp-2; othrws nd procssng. Stp-4: n a rcognton sgmnt mrg all canddats wth th sam Chns charactr nto on sngl canddat whos postror probablty quals to th sum of postror probablts of all canddats mrgd. Fg. 6(b) shows charactr canddats gnratd by splttng th algnd ntwork of Fg. 6(a). As can b sn all th word hypothss splts nto Chns charactrs and Chns charactrs wth compttv rlatonshp ar groupd nto th sam rcognton sgmnt. Fgur 6(a) gvs th algnd ntwork gnratd by algnng th Chns word lattc of Fgur 4(a) Algorthm for Charactr Canddat Gnraton Basd on th algnd ntwork charactr canddat ar gnratd by splttng words nto charactrs and rordrng all canddats accordng to postror probablts. Th algorthm can b dscrbd as follows: Stp-1: ntalz: n=0 m =0. Stp-2: lt num = mn{ u ( )} u ( ) rmans th (a) f u = E n sam as bfor. For ach lnk n 1 2. ( ) num lt th word of canddat c b W c = QW ( ) and th postror probablty of canddat c b Pc = p ( ) st C C c 01 num1 m m Fgur 6. Exampl of Chns canddat gnraton 4. Exprmntal rsults To tst th ffctvnss of th proposd canddat gnraton mthod for ntractv Chns spch rcognton spch rcognton and canddat gnraton ar conductd on a contnuous Chns spch corpus known as th 863 Corpus. Th corpus contans data of 80 spakrs and 520 uttrancs ar avalabl for ach spakr. Data of 50 spakrs ar usd as tranng st and 450 sntncs xtractd from th rst data ar usd for tstng. MFCC faturs and tr-phon 586 4

5 HMMs ar usd for acoustc modl and th languag modl s a b-gram modl wth 120k words trand on 25M Popl s Daly Corpus. To masur th rror corrcton prformanc by usng th proposd algorthm to gnrat Chns canddats covrag rat of corrct charactrs n th top N canddats ar usd as mtrcs whch s also rfrrd to as charactr corrct rat. Th covrag rats for dffrnt numbrs of canddats gnratd by th proposd algorthm ar shown n Tabl 1. Th Top-N rprsnts th N-bst canddats and th Top-1 has no compttv canddats xcpt th bst word. t can b sn from th tabl that word corrctnss was mprovd whn th numbr of canddats ncrasd. Th word corrctnss fnally achvd was 94.06% and most rronous words could b corrctd whn th numbr of canddats was tn. Ths rsults show that th canddat gnratd by th proposd algorthm s suffcntly ffcnt to nabl a usr to corrct rcognton rrors. Tabl 1. Word corrctnss wth dffrnt numbr of canddats Numbr of Canddats Top % Top % Top % Top % all 94.06% Charactr Corrct Rat Acknowldgmnt Ths work s sponsord by Natonal Ky Tchnology R&D Program of Chna (Grant No. 2008BAH26B03). Rfrncs [1] Suhm B. Myrs B. & Wabl A. Dsgnng ntractv rror Rcovry Mthods for Spch ntrfacs Procdngs of ACM CH 1996 Workshop on Dsgnng th Usr ntrfac for Spch Rcognton applcatons. [2] Brnhard Suhm Emprcal Evaluaton of ntractv Multmodal Error Corrcton Proc. EEE Workshop on Spch rcognton and Undrstandng [3] Karat C. Halvrson C. Horn D. and Karat Pattrns of Entry and Corrcton n Larg Vocabulary Contnuous Spch Rcognton Systms Proc. CH [4] Ogata J. Goto M Spch rpar: quck rror corrcton ust by usng slcton opraton for spch nput ntrfacs. n: Proc. ntrspch pp [5] L. Mangu E. Brll and A. Stolck Fndng consnsus n spch rcognton: word rror mnzaton and othr applcaton of confuson ntwork Computr Spch and Languag vol.14 (4) pp [6] L. Mangu Fndng Consnsus n Spch Rcognton PhD Thss Johns Hopkns Unvrsty [7] J. Xu and Y.-X. Zhao mprovd confuson ntwork algorthm and shortst path sarch from word lattc CASSP 2005 vol.1 pp Conclusons n ths papr th prncpls of canddat gnraton ar dscussd and a canddat gnraton algorthm for ntractv Chns spch rcognton s proposd. Unlk mthods for Englsh spch whch ar not sutabl for Chns spch th proposd mthod lsts all compttv canddats n forms of Chns charactrs n stad of words. Th man da s to algn tm-ovrlappd lnks n Chns word lattc nto clustrs basd on thr phontc smlarty and thn obtan canddats by splttng words nto Chns charactrs. Exprmntal rsults hav shown that th canddat gnratd by th proposd algorthm can b usd to corrct most rcognton rrors. n futur work bas on th Chns canddat gnratd by th proposd algorthm w ntnd to us mor ntractv nformaton to mprov th rror corrcton prformanc of ntractv Chns spch rcognton

6 Ths full txt papr was pr rvwd at th drcton of EEE Communcatons Socty subct mattr xprts for publcaton n th WCNC 2008 procdngs. Numbr of rsolvd nods Fg ML wth Out-of-Rang ML Numbr of nods n th ntwork Th ffct of nod dgr n locaton dscovry nods rmans all th sam n th scnaro of th ntwork wth 24 nods. Th rason bhnd s that nods mult-hop away may b gographcally too far away from th unknown nod U so that thy ar out of th rang of any possbl locaton U mght rsd at. Thrfor 2 or 3 can b usd as a practcal sttng of h. C. Effct of avrag nod dgr Th numbr of addtonal nods locatd usng Out-of-Rang nformaton dpnds on th ntwork connctvty and topology. Fg 8 prsnts th numbr of rsolvd nods aftr locaton dscovry wth Out-of-Rang and wthout Out-of-Rang nformaton n a st of scnaros. Th total numbr of nods n th ntwork ncrass from 16 to 48 whl th numbr of rfrnc nods rmans to b 4. Th avrag nod dgr n ach scnaro s lstd n Tabl. t s worth notng that th avrag nod dgr dos not always ncras as th numbr of nods ncrass n th ntwork. Although aganst ntuton a furthr look at th topology shows that t could actually occur. t stms from th fact that nods ar unformly placd ovr th ntr ara. n th unform placmnt th ara s dvdd nto a numbr of clls and nods ar randomly placd wthn ach cll. A slght ncrasng of th numbr of nods can rsult n an addtonal cll wth only fw nods placd n t. TABLE AVERAGE NODE DEGREE g N A NETWORK OF n NODES n g Whn th ntwork has a total of 16 nods no othr nod can b locatd by mult-latraton or our schm snc th ntwork s parttond. No Out-of-Rang nformaton s usful n ths scnaro du to th lack of connctvty. Our schm rsolvs mor nods locaton than th basc mult-latraton schm as th connctvty ncrass. As th ntwork bcoms hghly connctd almost all nods can rsolv thr locaton usng th basc mult-latraton schm. As a rsult no Out-of-Rang nformaton s ndd n ths cass. Thrfor th proposd schm can rsolv a larg numbr of addtonal snsor nods locaton usng Out-of-Rang whn th ntwork s sparsly yt suffcntly connctd. V. CONCLUSON AND FUTURE WORK n mult-latraton basd localzaton schms for snsor ntworks ach unknown nod must gan suffcnt nformaton such as thr rfrnc nods n ordr to dscovr ts locaton. n undr watr snsor ntworks howvr t s not uncommon that an unknown snsor nod dos not hav thr or mor rfrnc nghborng nods du to th spars ntwork topology and lmtd avalablty of rfrnc nods. As a rsult many nods can not dscovr thr locatons usng th mult-latraton schm. n ths papr w propos to utlz th Out-of-Rang nformaton at both rfrnc nods and unknown nods to dtrmn snsor nods locaton n UWSNs whn suffcnt nformaton s not avalabl. t s shown that an unknown nod wth only two nghborng rsolvd nods can dtrmn ts locaton n crtan scnaros. Th smulaton rsults show that th proposd schm can sgnfcantly ncras th numbr of rsolvd snsor nods aftr locaton dscovry complts whn th ntwork connctvty s low. Th ncras can b as bg as 50% n som scnaros. W focusd on locatng mor nods n UWSN usng Out-of- Rang nformaton n ths papr. Th proposd schm howvr has th potntal to b appld to som othr ntworks such as ad hoc snsor ntworks. Ths ssus wll b furthr nvstgatd n th futur. ACKNOWLEDGMENTS Ths work s supportd by NSF awards and REFERENCES [1]. F. Akyldz D. Pompl and T. Mloda Undrwatr acoustc snsor ntworks: rsarch challngs Ad Hoc Ntworks (Elsvr) pp [2] J. Aspns T. Ern D. K. Goldnbrg A. S. Mors W. Whtly Y. R. Yang B. D. Andrson and P. N. Blhumur A thory of ntwork localzaton EEE Transactons on Mobl Computng vol. 5 no. 12 pp [3] J. Gbson d. Th Mobl Communcatons Handbook. EEE Prss [4] D. K. Goldnbrg P. Bhlr M. Cao J. Fang B. D. Andrson A. S. Mors and Y. R. Yang Localzaton n Spars Ntworks usng Swps n MobCom 06: Procdngs of th 12th annual ntrnatonal confrnc on Mobl computng and ntworkng (Nw York NY USA) ACM Prss Sptmbr [5] D. K. Goldnbrg A. Krshnamurthy W. C. Manss Y. R. Yang A. Young A. S. Mors A. Savvds and B. D. Andrson Ntwork Localzaton n Partally Localzabl Ntworks n Procdngs of N- FOCOM 2005 (Mam FL) March [6] J. Hdmann W. Y J. Wlls A. Syd and Y. L Rsarch Challngs and Applcatons for Undrwatr Snsor Ntworkng n Procdngs of th EEE Wrlss Communcatons and Ntworkng Confrnc (Las Vgas Nvada USA) pp EEE Aprl [7] N. Malhotra M. Krasnwsk C. Yang S. Bagch and W. Chappll Locaton Estmaton n Ad-Hoc Ntworks wth Drctonal Antnnas n Procdngs of th 25th EEE ntrnatonal Confrnc on Dstrbutd Computng Systms (CDCS) (Columbus Oho USA.) Jun [8] D. Nculscu and B. Nath Ad hoc postonng systm (APS) n Procdngs of EEE Global Communcatons Confrnc (GLOBECOM) [9] J. Rc SEAWEB ACOUSTC COMMUNCATON AND NAVGA- TON NETWORKS n Procdngs of th ntrnatonal Confrnc Undrwatr Acoustc Masurmnts: Tchnologs and Rsults (Hraklon Crt Grc) Jun-July [10] A. Savvds C. chh Han and M. B. Strvastava ynamc Fn-Grand Localzaton n Ad-Hoc Ntworks of Snsors n Procdngs of th ACM ntrnatonal Confrnc on Mobl Computng and Ntworkng (MOB- COM01) (Rom taly) July [11] A. Savvds H. Park and M. B. Srvastava Th bts and flops of th n-hop multlatraton prmtv for nod localzaton problms n Procdngs of th 1st ACM ntrnatonal workshop on Wrlss snsor ntworks and applcatons (WSNA) (Nw York NY USA) pp ACM Prss [12] G. Zhou T. H S. Krshnamurthy and J. A. Stankovc mpact of Rado rrgularty on Wrlss Snsor Ntworks n Procdngs of th Scond ntrnatonal Confrnc on Mobl Systms Applcatons and Srvcs (MobSys) Jun Authorzd lcnsd us lmtd to: Tamkang Unvrsty. Downloadd on July at 05:32 from EEE Xplor. Rstrctons apply.

10/7/14. Mixture Models. Comp 135 Introduction to Machine Learning and Data Mining. Maximum likelihood estimation. Mixture of Normals in 1D

10/7/14. Mixture Models. Comp 135 Introduction to Machine Learning and Data Mining. Maximum likelihood estimation. Mixture of Normals in 1D Comp 35 Introducton to Machn Larnng and Data Mnng Fall 204 rofssor: Ron Khardon Mxtur Modls Motvatd by soft k-mans w dvlopd a gnratv modl for clustrng. Assum thr ar k clustrs Clustrs ar not rqurd to hav

More information

Soft k-means Clustering. Comp 135 Machine Learning Computer Science Tufts University. Mixture Models. Mixture of Normals in 1D

Soft k-means Clustering. Comp 135 Machine Learning Computer Science Tufts University. Mixture Models. Mixture of Normals in 1D Comp 35 Machn Larnng Computr Scnc Tufts Unvrsty Fall 207 Ron Khardon Th EM Algorthm Mxtur Modls Sm-Suprvsd Larnng Soft k-mans Clustrng ck k clustr cntrs : Assocat xampls wth cntrs p,j ~~ smlarty b/w cntr

More information

Outlier-tolerant parameter estimation

Outlier-tolerant parameter estimation Outlr-tolrant paramtr stmaton Baysan thods n physcs statstcs machn larnng and sgnal procssng (SS 003 Frdrch Fraundorfr fraunfr@cg.tu-graz.ac.at Computr Graphcs and Vson Graz Unvrsty of Tchnology Outln

More information

ON THE COMPLEXITY OF K-STEP AND K-HOP DOMINATING SETS IN GRAPHS

ON THE COMPLEXITY OF K-STEP AND K-HOP DOMINATING SETS IN GRAPHS MATEMATICA MONTISNIRI Vol XL (2017) MATEMATICS ON TE COMPLEXITY OF K-STEP AN K-OP OMINATIN SETS IN RAPS M FARAI JALALVAN AN N JAFARI RA partmnt of Mathmatcs Shahrood Unrsty of Tchnology Shahrood Iran Emals:

More information

Analyzing Frequencies

Analyzing Frequencies Frquncy (# ndvduals) Frquncy (# ndvduals) /3/16 H o : No dffrnc n obsrvd sz frquncs and that prdctd by growth modl How would you analyz ths data? 15 Obsrvd Numbr 15 Expctd Numbr from growth modl 1 1 5

More information

Review - Probabilistic Classification

Review - Probabilistic Classification Mmoral Unvrsty of wfoundland Pattrn Rcognton Lctur 8 May 5, 6 http://www.ngr.mun.ca/~charlsr Offc Hours: Tusdays Thursdays 8:3-9:3 PM E- (untl furthr notc) Gvn lablld sampls { ɛc,,,..., } {. Estmat Rvw

More information

The Hyperelastic material is examined in this section.

The Hyperelastic material is examined in this section. 4. Hyprlastcty h Hyprlastc matral s xad n ths scton. 4..1 Consttutv Equatons h rat of chang of ntrnal nrgy W pr unt rfrnc volum s gvn by th strss powr, whch can b xprssd n a numbr of dffrnt ways (s 3.7.6):

More information

Lucas Test is based on Euler s theorem which states that if n is any integer and a is coprime to n, then a φ(n) 1modn.

Lucas Test is based on Euler s theorem which states that if n is any integer and a is coprime to n, then a φ(n) 1modn. Modul 10 Addtonal Topcs 10.1 Lctur 1 Prambl: Dtrmnng whthr a gvn ntgr s prm or compost s known as prmalty tstng. Thr ar prmalty tsts whch mrly tll us whthr a gvn ntgr s prm or not, wthout gvng us th factors

More information

ST 524 NCSU - Fall 2008 One way Analysis of variance Variances not homogeneous

ST 524 NCSU - Fall 2008 One way Analysis of variance Variances not homogeneous ST 54 NCSU - Fall 008 On way Analyss of varanc Varancs not homognous On way Analyss of varanc Exampl (Yandll, 997) A plant scntst masurd th concntraton of a partcular vrus n plant sap usng ELISA (nzym-lnkd

More information

Journal of Theoretical and Applied Information Technology 10 th January Vol. 47 No JATIT & LLS. All rights reserved.

Journal of Theoretical and Applied Information Technology 10 th January Vol. 47 No JATIT & LLS. All rights reserved. Journal o Thortcal and Appld Inormaton Tchnology th January 3. Vol. 47 No. 5-3 JATIT & LLS. All rghts rsrvd. ISSN: 99-8645 www.att.org E-ISSN: 87-395 RESEARCH ON PROPERTIES OF E-PARTIAL DERIVATIVE OF LOGIC

More information

Economics 600: August, 2007 Dynamic Part: Problem Set 5. Problems on Differential Equations and Continuous Time Optimization

Economics 600: August, 2007 Dynamic Part: Problem Set 5. Problems on Differential Equations and Continuous Time Optimization THE UNIVERSITY OF MARYLAND COLLEGE PARK, MARYLAND Economcs 600: August, 007 Dynamc Part: Problm St 5 Problms on Dffrntal Equatons and Contnuous Tm Optmzaton Quston Solv th followng two dffrntal quatons.

More information

CHAPTER 33: PARTICLE PHYSICS

CHAPTER 33: PARTICLE PHYSICS Collg Physcs Studnt s Manual Chaptr 33 CHAPTER 33: PARTICLE PHYSICS 33. THE FOUR BASIC FORCES 4. (a) Fnd th rato of th strngths of th wak and lctromagntc forcs undr ordnary crcumstancs. (b) What dos that

More information

An Overview of Markov Random Field and Application to Texture Segmentation

An Overview of Markov Random Field and Application to Texture Segmentation An Ovrvw o Markov Random Fld and Applcaton to Txtur Sgmntaton Song-Wook Joo Octobr 003. What s MRF? MRF s an xtnson o Markov Procss MP (D squnc o r.v. s unlatral (causal: p(x t x,

More information

A Note on Estimability in Linear Models

A Note on Estimability in Linear Models Intrnatonal Journal of Statstcs and Applcatons 2014, 4(4): 212-216 DOI: 10.5923/j.statstcs.20140404.06 A Not on Estmablty n Lnar Modls S. O. Adymo 1,*, F. N. Nwob 2 1 Dpartmnt of Mathmatcs and Statstcs,

More information

Grand Canonical Ensemble

Grand Canonical Ensemble Th nsmbl of systms mmrsd n a partcl-hat rsrvor at constant tmpratur T, prssur P, and chmcal potntal. Consdr an nsmbl of M dntcal systms (M =,, 3,...M).. Thy ar mutually sharng th total numbr of partcls

More information

Chapter 6 Student Lecture Notes 6-1

Chapter 6 Student Lecture Notes 6-1 Chaptr 6 Studnt Lctur Nots 6-1 Chaptr Goals QM353: Busnss Statstcs Chaptr 6 Goodnss-of-Ft Tsts and Contngncy Analyss Aftr compltng ths chaptr, you should b abl to: Us th ch-squar goodnss-of-ft tst to dtrmn

More information

Authentication Transmission Overhead Between Entities in Mobile Networks

Authentication Transmission Overhead Between Entities in Mobile Networks 0 IJCSS Intrnatonal Journal of Computr Scnc and twork Scurty, VO.6 o.b, March 2006 Authntcaton Transmsson Ovrhad Btwn Entts n Mobl tworks Ja afr A-Sararh and Sufan Yousf Faculty of Scnc and Tchnology,

More information

Naresuan University Journal: Science and Technology 2018; (26)1

Naresuan University Journal: Science and Technology 2018; (26)1 Narsuan Unvrsty Journal: Scnc and Tchnology 018; (6)1 Th Dvlopmnt o a Corrcton Mthod or Ensurng a Contnuty Valu o Th Ch-squar Tst wth a Small Expctd Cll Frquncy Kajta Matchma 1 *, Jumlong Vongprasrt and

More information

COMPLEX NUMBER PAIRWISE COMPARISON AND COMPLEX NUMBER AHP

COMPLEX NUMBER PAIRWISE COMPARISON AND COMPLEX NUMBER AHP ISAHP 00, Bal, Indonsa, August -9, 00 COMPLEX NUMBER PAIRWISE COMPARISON AND COMPLEX NUMBER AHP Chkako MIYAKE, Kkch OHSAWA, Masahro KITO, and Masaak SHINOHARA Dpartmnt of Mathmatcal Informaton Engnrng

More information

??? Dynamic Causal Modelling for M/EEG. Electroencephalography (EEG) Dynamic Causal Modelling. M/EEG analysis at sensor level. time.

??? Dynamic Causal Modelling for M/EEG. Electroencephalography (EEG) Dynamic Causal Modelling. M/EEG analysis at sensor level. time. Elctroncphalography EEG Dynamc Causal Modllng for M/EEG ampltud μv tm ms tral typ 1 tm channls channls tral typ 2 C. Phllps, Cntr d Rchrchs du Cyclotron, ULg, Blgum Basd on slds from: S. Kbl M/EEG analyss

More information

Folding of Regular CW-Complexes

Folding of Regular CW-Complexes Ald Mathmatcal Scncs, Vol. 6,, no. 83, 437-446 Foldng of Rgular CW-Comlxs E. M. El-Kholy and S N. Daoud,3. Dartmnt of Mathmatcs, Faculty of Scnc Tanta Unvrsty,Tanta,Egyt. Dartmnt of Mathmatcs, Faculty

More information

Polytropic Process. A polytropic process is a quasiequilibrium process described by

Polytropic Process. A polytropic process is a quasiequilibrium process described by Polytropc Procss A polytropc procss s a quasqulbrum procss dscrbd by pv n = constant (Eq. 3.5 Th xponnt, n, may tak on any valu from to dpndng on th partcular procss. For any gas (or lqud, whn n = 0, th

More information

te Finance (4th Edition), July 2017.

te Finance (4th Edition), July 2017. Appndx Chaptr. Tchncal Background Gnral Mathmatcal and Statstcal Background Fndng a bas: 3 2 = 9 3 = 9 1 /2 x a = b x = b 1/a A powr of 1 / 2 s also quvalnt to th squar root opraton. Fndng an xponnt: 3

More information

Abstract Interpretation: concrete and abstract semantics

Abstract Interpretation: concrete and abstract semantics Abstract Intrprtation: concrt and abstract smantics Concrt smantics W considr a vry tiny languag that manags arithmtic oprations on intgrs valus. Th (concrt) smantics of th languags cab b dfind by th funzcion

More information

Minimizing Energy Consumption in Wireless Ad hoc Networks with Meta heuristics

Minimizing Energy Consumption in Wireless Ad hoc Networks with Meta heuristics Avalabl onln at www.scncdrct.com Procda Computr Scnc 19 (2013 ) 106 115 Th 4 th Intrnatonal Confrnc on Ambnt Systms, Ntworks and Tchnologs (ANT 2013) Mnmzng Enrgy Consumpton n Wrlss Ad hoc Ntworks wth

More information

Search sequence databases 3 10/25/2016

Search sequence databases 3 10/25/2016 Sarch squnc databass 3 10/25/2016 Etrm valu distribution Ø Suppos X is a random variabl with probability dnsity function p(, w sampl a larg numbr S of indpndnt valus of X from this distribution for an

More information

Lecture 3: Phasor notation, Transfer Functions. Context

Lecture 3: Phasor notation, Transfer Functions. Context EECS 5 Fall 4, ctur 3 ctur 3: Phasor notaton, Transfr Functons EECS 5 Fall 3, ctur 3 Contxt In th last lctur, w dscussd: how to convrt a lnar crcut nto a st of dffrntal quatons, How to convrt th st of

More information

1 Minimum Cut Problem

1 Minimum Cut Problem CS 6 Lctur 6 Min Cut and argr s Algorithm Scribs: Png Hui How (05), Virginia Dat: May 4, 06 Minimum Cut Problm Today, w introduc th minimum cut problm. This problm has many motivations, on of which coms

More information

Preview. Graph. Graph. Graph. Graph Representation. Graph Representation 12/3/2018. Graph Graph Representation Graph Search Algorithms

Preview. Graph. Graph. Graph. Graph Representation. Graph Representation 12/3/2018. Graph Graph Representation Graph Search Algorithms /3/0 Prvw Grph Grph Rprsntton Grph Srch Algorthms Brdth Frst Srch Corrctnss of BFS Dpth Frst Srch Mnmum Spnnng Tr Kruskl s lgorthm Grph Drctd grph (or dgrph) G = (V, E) V: St of vrt (nod) E: St of dgs

More information

A Packet Buffer Evaluation Method Exploiting Queueing Theory for Wireless Sensor Networks

A Packet Buffer Evaluation Method Exploiting Queueing Theory for Wireless Sensor Networks DOI: 10.2298/CSIS110227057Q A Packt Buffr Evaluaton Mthod Explotng Quung Thory for Wrlss Snsor Ntworks T Qu 1,2, Ln Fng 2, Fng Xa 1*, Guow Wu 1, and Yu Zhou 1 1 School of Softwar, Dalan Unvrsty of Tchnology,

More information

Physics of Very High Frequency (VHF) Capacitively Coupled Plasma Discharges

Physics of Very High Frequency (VHF) Capacitively Coupled Plasma Discharges Physcs of Vry Hgh Frquncy (VHF) Capactvly Coupld Plasma Dschargs Shahd Rauf, Kallol Bra, Stv Shannon, and Kn Collns Appld Matrals, Inc., Sunnyval, CA AVS 54 th Intrnatonal Symposum Sattl, WA Octobr 15-19,

More information

Optimal Ordering Policy in a Two-Level Supply Chain with Budget Constraint

Optimal Ordering Policy in a Two-Level Supply Chain with Budget Constraint Optmal Ordrng Polcy n a Two-Lvl Supply Chan wth Budgt Constrant Rasoul aj Alrza aj Babak aj ABSTRACT Ths papr consdrs a two- lvl supply chan whch consst of a vndor and svral rtalrs. Unsatsfd dmands n rtalrs

More information

A New Competitive Ratio for Network Applications with Hard Performance Guarantees

A New Competitive Ratio for Network Applications with Hard Performance Guarantees A Nw Compttv Rato for Ntwork Applcatons wth Hard Prformanc Guarants Han Dng Dpartmnt of ECE Txas A&M Unvrsty Collg Staton, TX 77840, USA Emal: hdng@maltamudu I-Hong Hou Dpartmnt of ECE Txas A&M Unvrsty

More information

Group Consensus of Multi-agent Networks With Multiple Time Delays

Group Consensus of Multi-agent Networks With Multiple Time Delays Intrnatonal Confrnc on Artfcal Intllgnc: Tchnologs and Applcatons (ICAITA 06 Group Consnsus of Mult-agnt Ntworks Wth Multpl Tm Dlays Langhao J* Xnyu Zhao Qun Lu and Yong Wang Chongqng Ky Laboratory of

More information

Higher order derivatives

Higher order derivatives Robrto s Nots on Diffrntial Calculus Chaptr 4: Basic diffrntiation ruls Sction 7 Highr ordr drivativs What you nd to know alrady: Basic diffrntiation ruls. What you can larn hr: How to rpat th procss of

More information

Group Codes Define Over Dihedral Groups of Small Order

Group Codes Define Over Dihedral Groups of Small Order Malaysan Journal of Mathmatcal Scncs 7(S): 0- (0) Spcal Issu: Th rd Intrnatonal Confrnc on Cryptology & Computr Scurty 0 (CRYPTOLOGY0) MALAYSIA JOURAL OF MATHEMATICAL SCIECES Journal hompag: http://nspm.upm.du.my/ournal

More information

8-node quadrilateral element. Numerical integration

8-node quadrilateral element. Numerical integration Fnt Elmnt Mthod lctur nots _nod quadrlatral lmnt Pag of 0 -nod quadrlatral lmnt. Numrcal ntgraton h tchnqu usd for th formulaton of th lnar trangl can b formall tndd to construct quadrlatral lmnts as wll

More information

Decentralized Adaptive Control and the Possibility of Utilization of Networked Control System

Decentralized Adaptive Control and the Possibility of Utilization of Networked Control System Dcntralzd Adaptv Control and th Possblty of Utlzaton of Ntworkd Control Systm MARIÁN ÁRNÍK, JÁN MURGAŠ Slovak Unvrsty of chnology n Bratslava Faculty of Elctrcal Engnrng and Informaton chnology Insttut

More information

Three-Node Euler-Bernoulli Beam Element Based on Positional FEM

Three-Node Euler-Bernoulli Beam Element Based on Positional FEM Avalabl onln at www.scncdrct.com Procda Engnrng 9 () 373 377 Intrnatonal Workshop on Informaton and Elctroncs Engnrng (IWIEE) Thr-Nod Eulr-Brnoull Bam Elmnt Basd on Postonal FEM Lu Jan a *,b, Zhou Shnj

More information

Highly Imperceptible and Reversible Text Steganography Using Invisible Character based Codeword

Highly Imperceptible and Reversible Text Steganography Using Invisible Character based Codeword Assocaton for Informaton Systms AIS Elctronc Lbrary (AISL) PACIS 2017 Procdngs Pacfc Asa Confrnc on Informaton Systms (PACIS) Summr 7-19-2017 Hghly Imprcptbl and Rvrsbl Txt Stganography Usng Invsbl Charactr

More information

SCITECH Volume 5, Issue 1 RESEARCH ORGANISATION November 17, 2015

SCITECH Volume 5, Issue 1 RESEARCH ORGANISATION November 17, 2015 Journal of Informaton Scncs and Computng Tchnologs(JISCT) ISSN: 394-966 SCITECH Volum 5, Issu RESEARCH ORGANISATION Novmbr 7, 5 Journal of Informaton Scncs and Computng Tchnologs www.sctcrsarch.com/journals

More information

Quantum-Inspired Bee Colony Algorithm

Quantum-Inspired Bee Colony Algorithm Opn Journal of Optmzaton, 05, 4, 5-60 Publshd Onln Sptmbr 05 n ScRs. http://www.scrp.org/ournal/oop http://dx.do.org/0.436/oop.05.43007 Quantum-Insprd B Colony Algorthm Guoru L, Mu Sun, Panch L School

More information

Logistic Regression I. HRP 261 2/10/ am

Logistic Regression I. HRP 261 2/10/ am Logstc Rgrsson I HRP 26 2/0/03 0- am Outln Introducton/rvw Th smplst logstc rgrsson from a 2x2 tabl llustrats how th math works Stp-by-stp xampls to b contnud nxt tm Dummy varabls Confoundng and ntracton

More information

Jones vector & matrices

Jones vector & matrices Jons vctor & matrcs PY3 Colást na hollscol Corcagh, Ér Unvrst Collg Cork, Irland Dpartmnt of Phscs Matr tratmnt of polarzaton Consdr a lght ra wth an nstantanous -vctor as shown k, t ˆ k, t ˆ k t, o o

More information

Lecture 14. Relic neutrinos Temperature at neutrino decoupling and today Effective degeneracy factor Neutrino mass limits Saha equation

Lecture 14. Relic neutrinos Temperature at neutrino decoupling and today Effective degeneracy factor Neutrino mass limits Saha equation Lctur Rlc nutrnos mpratur at nutrno dcoupln and today Effctv dnracy factor Nutrno mass lmts Saha quaton Physcal Cosmoloy Lnt 005 Rlc Nutrnos Nutrnos ar wakly ntractn partcls (lptons),,,,,,, typcal ractons

More information

2. Grundlegende Verfahren zur Übertragung digitaler Signale (Zusammenfassung) Informationstechnik Universität Ulm

2. Grundlegende Verfahren zur Übertragung digitaler Signale (Zusammenfassung) Informationstechnik Universität Ulm . Grundlgnd Vrfahrn zur Übrtragung dgtalr Sgnal (Zusammnfassung) wt Dc. 5 Transmsson of Dgtal Sourc Sgnals Sourc COD SC COD MOD MOD CC dg RF s rado transmsson mdum Snk DC SC DC CC DM dg DM RF g physcal

More information

Discrete Shells Simulation

Discrete Shells Simulation Dscrt Shlls Smulaton Xaofng M hs proct s an mplmntaton of Grnspun s dscrt shlls, th modl of whch s govrnd by nonlnar mmbran and flxural nrgs. hs nrgs masur dffrncs btwns th undformd confguraton and th

More information

Minimum Spanning Trees

Minimum Spanning Trees Mnmum Spnnng Trs Spnnng Tr A tr (.., connctd, cyclc grph) whch contns ll th vrtcs of th grph Mnmum Spnnng Tr Spnnng tr wth th mnmum sum of wghts 1 1 Spnnng forst If grph s not connctd, thn thr s spnnng

More information

Optimal Data Transmission and Channel Code Rate Allocation in Multi-path Wireless Networks

Optimal Data Transmission and Channel Code Rate Allocation in Multi-path Wireless Networks Optmal Data Transmsson and Channl Cod Rat Allocaton n Mult-path Wrlss Ntwors Kvan Ronas, Amr-Hamd Mohsnan-Rad,VncntW.S.Wong, Sathsh Gopalarshnan, and Robrt Schobr Dpartmnt of Elctrcal and Computr Engnrng

More information

Fakultät III Univ.-Prof. Dr. Jan Franke-Viebach

Fakultät III Univ.-Prof. Dr. Jan Franke-Viebach Unv.Prof. r. J. FrankVbach WS 067: Intrnatonal Economcs ( st xam prod) Unvrstät Sgn Fakultät III Unv.Prof. r. Jan FrankVbach Exam Intrnatonal Economcs Wntr Smstr 067 ( st Exam Prod) Avalabl tm: 60 mnuts

More information

A NEW GENERALISATION OF SAM-SOLAI S MULTIVARIATE ADDITIVE GAMMA DISTRIBUTION*

A NEW GENERALISATION OF SAM-SOLAI S MULTIVARIATE ADDITIVE GAMMA DISTRIBUTION* A NEW GENERALISATION OF SAM-SOLAI S MULTIVARIATE ADDITIVE GAMMA DISTRIBUTION* Dr. G.S. Davd Sam Jayakumar, Assstant Profssor, Jamal Insttut of Managmnt, Jamal Mohamd Collg, Truchraall 620 020, South Inda,

More information

September 27, Introduction to Ordinary Differential Equations. ME 501A Seminar in Engineering Analysis Page 1. Outline

September 27, Introduction to Ordinary Differential Equations. ME 501A Seminar in Engineering Analysis Page 1. Outline Introucton to Ornar Dffrntal Equatons Sptmbr 7, 7 Introucton to Ornar Dffrntal Equatons Larr artto Mchancal Engnrng AB Smnar n Engnrng Analss Sptmbr 7, 7 Outln Rvw numrcal solutons Bascs of ffrntal quatons

More information

Emotion Recognition from Speech Using IG-Based Feature Compensation

Emotion Recognition from Speech Using IG-Based Feature Compensation Computatonal Lngustcs and Chns Languag Procssng Vol. 12, No. 1, March 2007, pp. 65-78 65 Th Assocaton for Computatonal Lngustcs and Chns Languag Procssng Emoton Rcognton from Spch Usng IG-Basd Fatur Compnsaton

More information

HORIZONTAL IMPEDANCE FUNCTION OF SINGLE PILE IN SOIL LAYER WITH VARIABLE PROPERTIES

HORIZONTAL IMPEDANCE FUNCTION OF SINGLE PILE IN SOIL LAYER WITH VARIABLE PROPERTIES 13 th World Confrnc on Earthquak Engnrng Vancouvr, B.C., Canada August 1-6, 4 Papr No. 485 ORIZONTAL IMPEDANCE FUNCTION OF SINGLE PILE IN SOIL LAYER WIT VARIABLE PROPERTIES Mngln Lou 1 and Wnan Wang Abstract:

More information

Learning Spherical Convolution for Fast Features from 360 Imagery

Learning Spherical Convolution for Fast Features from 360 Imagery Larning Sphrical Convolution for Fast Faturs from 36 Imagry Anonymous Author(s) 3 4 5 6 7 8 9 3 4 5 6 7 8 9 3 4 5 6 7 8 9 3 3 3 33 34 35 In this fil w provid additional dtails to supplmnt th main papr

More information

External Equivalent. EE 521 Analysis of Power Systems. Chen-Ching Liu, Boeing Distinguished Professor Washington State University

External Equivalent. EE 521 Analysis of Power Systems. Chen-Ching Liu, Boeing Distinguished Professor Washington State University xtrnal quvalnt 5 Analyss of Powr Systms Chn-Chng Lu, ong Dstngushd Profssor Washngton Stat Unvrsty XTRNAL UALNT ach powr systm (ara) s part of an ntrconnctd systm. Montorng dvcs ar nstalld and data ar

More information

CHAPTER 7d. DIFFERENTIATION AND INTEGRATION

CHAPTER 7d. DIFFERENTIATION AND INTEGRATION CHAPTER 7d. DIFFERENTIATION AND INTEGRATION A. J. Clark School o Engnrng Dpartmnt o Cvl and Envronmntal Engnrng by Dr. Ibrahm A. Assakka Sprng ENCE - Computaton Mthods n Cvl Engnrng II Dpartmnt o Cvl and

More information

SENSOR networks are wireless ad hoc networks used for. Minimum Energy Fault Tolerant Sensor Networks

SENSOR networks are wireless ad hoc networks used for. Minimum Energy Fault Tolerant Sensor Networks Mnmum Enrgy Fault Tolrant Snsor Ntworks Ptar Djukc and Shahrokh Vala Th Edward S Rogrs Sr Dpartmnt of Elctrcal and Computr Engnrng Unvrsty of Toronto, 0 Kng s Collg Road, Toronto, ON, MS 3G4, Canada -mal:{djukc,vala}@commutorontoca

More information

The Equitable Dominating Graph

The Equitable Dominating Graph Intrnational Journal of Enginring Rsarch and Tchnology. ISSN 0974-3154 Volum 8, Numbr 1 (015), pp. 35-4 Intrnational Rsarch Publication Hous http://www.irphous.com Th Equitabl Dominating Graph P.N. Vinay

More information

From Structural Analysis to FEM. Dhiman Basu

From Structural Analysis to FEM. Dhiman Basu From Structural Analyss to FEM Dhman Basu Acknowldgmnt Followng txt books wr consultd whl prparng ths lctur nots: Znkwcz, OC O.C. andtaylor Taylor, R.L. (000). Th FntElmnt Mthod, Vol. : Th Bass, Ffth dton,

More information

A Robust Fuzzy Support Vector Machine for Two-class Pattern Classification

A Robust Fuzzy Support Vector Machine for Two-class Pattern Classification 76 Intrnatonal Journal of Fuzzy Systms, Vol. 8, o., Jun 006 A Robust Fuzzy Support Vctor Machn for wo-class Pattrn Classfcaton G. H. L, J. S. aur, and C.W. ao Abstract hs papr proposs a systmatc mthod

More information

The following manuscript was published in

The following manuscript was published in h followng manuscrpt was publshd n JunW Hsh Shang-L Yu and Yung-Shng Chn Moton-basd vdo rtrval by trajctory matchng IEEE ransactons on Crcuts and Systms for Vdo chnology Vol. 16 No. 3 396-409 2006. Moton-basd

More information

Computing and Communications -- Network Coding

Computing and Communications -- Network Coding 89 90 98 00 Computing and Communications -- Ntwork Coding Dr. Zhiyong Chn Institut of Wirlss Communications Tchnology Shanghai Jiao Tong Univrsity China Lctur 5- Nov. 05 0 Classical Information Thory Sourc

More information

Fakultät III Wirtschaftswissenschaften Univ.-Prof. Dr. Jan Franke-Viebach

Fakultät III Wirtschaftswissenschaften Univ.-Prof. Dr. Jan Franke-Viebach Unvrstät Sgn Fakultät III Wrtschaftswssnschaftn Unv.-rof. Dr. Jan Frank-Vbach Exam Intrnatonal Fnancal Markts Summr Smstr 206 (2 nd Exam rod) Avalabl tm: 45 mnuts Soluton For your attnton:. las do not

More information

Binary Decision Diagram with Minimum Expected Path Length

Binary Decision Diagram with Minimum Expected Path Length Bnary Dcson Dagram wth Mnmum Expctd Path Lngth Y-Yu Lu Kuo-Hua Wang TngTng Hwang C. L. Lu Dpartmnt of Computr Scnc, Natonal Tsng Hua Unvrsty, Hsnchu 300, Tawan Dpt. of Computr Scnc and Informaton Engnrng,

More information

On the Capacity-Performance Trade-off of Online Policy in Delayed Mobile Offloading

On the Capacity-Performance Trade-off of Online Policy in Delayed Mobile Offloading On th Capacty-Prformanc Trad-off of Onln Polcy n Dlayd Mobl Offloadng Han Dng and I-Hong Hou Abstract WF offloadng, whr mobl usrs opportunstcally obtan data through WF rathr than cllular ntworks, s a promsng

More information

An Efficient Approach Based on Neuro-Fuzzy for Phishing Detection

An Efficient Approach Based on Neuro-Fuzzy for Phishing Detection Journal of Automaton and Control Engnrng Vol. 4, No. 2, Aprl 206 An Effcnt Approach Basd on Nuro-Fuzzy for Phshng Dtcton Luong Anh Tuan Nguyn, Huu Khuong Nguyn, and Ba Lam To Ho Ch Mnh Cty Unvrsty of Transport,

More information

Phy213: General Physics III 4/10/2008 Chapter 22 Worksheet 1. d = 0.1 m

Phy213: General Physics III 4/10/2008 Chapter 22 Worksheet 1. d = 0.1 m hy3: Gnral hyscs III 4/0/008 haptr Worksht lctrc Flds: onsdr a fxd pont charg of 0 µ (q ) q = 0 µ d = 0 a What s th agntud and drcton of th lctrc fld at a pont, a dstanc of 0? q = = 8x0 ˆ o d ˆ 6 N ( )

More information

Derangements and Applications

Derangements and Applications 2 3 47 6 23 Journal of Intgr Squncs, Vol. 6 (2003), Articl 03..2 Drangmnts and Applications Mhdi Hassani Dpartmnt of Mathmatics Institut for Advancd Studis in Basic Scincs Zanjan, Iran mhassani@iasbs.ac.ir

More information

Heisenberg Model. Sayed Mohammad Mahdi Sadrnezhaad. Supervisor: Prof. Abdollah Langari

Heisenberg Model. Sayed Mohammad Mahdi Sadrnezhaad. Supervisor: Prof. Abdollah Langari snbrg Modl Sad Mohammad Mahd Sadrnhaad Survsor: Prof. bdollah Langar bstract: n ths rsarch w tr to calculat analtcall gnvalus and gnvctors of fnt chan wth ½-sn artcls snbrg modl. W drov gnfuctons for closd

More information

Maneuvering Target Tracking Using Current Statistical Model Based Adaptive UKF for Wireless Sensor Network

Maneuvering Target Tracking Using Current Statistical Model Based Adaptive UKF for Wireless Sensor Network Journal of Communcatons Vol. 0, No. 8, August 05 Manuvrng argt racng Usng Currnt Statstcal Modl Basd Adaptv UKF for Wrlss Snsor Ntwor Xaojun Png,, Kuntao Yang, and Chang Lu School of Optcal and Elctronc

More information

MUSIC Based on Uniform Circular Array and Its Direction Finding Efficiency

MUSIC Based on Uniform Circular Array and Its Direction Finding Efficiency Intrnatonal Journal of Sgnal Procssng Systms Vol. 1, No. 2 Dcmbr 2013 MUSIC Basd on Unform Crcular Array and Its Drcton Fndng Effcncy Baofa Sun Dpartmnt of Computr Scnc and Tchnology, Anhu Sanlan Unvrsty,

More information

Maneuvering Target Tracking Using Current Statistical Model Based Adaptive UKF for Wireless Sensor Network

Maneuvering Target Tracking Using Current Statistical Model Based Adaptive UKF for Wireless Sensor Network Journal of Communcatons Vol. 0, No. 8, August 05 Manuvrng argt racng Usng Currnt Statstcal Modl Basd Adaptv UKF for Wrlss Snsor Ntwor Xaojun Png,, Kuntao Yang, and Chang Lu School of Optcal and Elctronc

More information

Epistemic Foundations of Game Theory. Lecture 1

Epistemic Foundations of Game Theory. Lecture 1 Royal Nthrlands cadmy of rts and Scncs (KNW) Mastr Class mstrdam, Fbruary 8th, 2007 Epstmc Foundatons of Gam Thory Lctur Gacomo onanno (http://www.con.ucdavs.du/faculty/bonanno/) QUESTION: What stratgs

More information

The Fourier Transform

The Fourier Transform /9/ Th ourr Transform Jan Baptst Josph ourr 768-83 Effcnt Data Rprsntaton Data can b rprsntd n many ways. Advantag usng an approprat rprsntaton. Eampls: osy ponts along a ln Color spac rd/grn/blu v.s.

More information

Observer Bias and Reliability By Xunchi Pu

Observer Bias and Reliability By Xunchi Pu Obsrvr Bias and Rliability By Xunchi Pu Introduction Clarly all masurmnts or obsrvations nd to b mad as accuratly as possibl and invstigators nd to pay carful attntion to chcking th rliability of thir

More information

Hardy-Littlewood Conjecture and Exceptional real Zero. JinHua Fei. ChangLing Company of Electronic Technology Baoji Shannxi P.R.

Hardy-Littlewood Conjecture and Exceptional real Zero. JinHua Fei. ChangLing Company of Electronic Technology Baoji Shannxi P.R. Hardy-Littlwood Conjctur and Excptional ral Zro JinHua Fi ChangLing Company of Elctronic Tchnology Baoji Shannxi P.R.China E-mail: fijinhuayoujian@msn.com Abstract. In this papr, w assum that Hardy-Littlwood

More information

Elements of Statistical Thermodynamics

Elements of Statistical Thermodynamics 24 Elmnts of Statistical Thrmodynamics Statistical thrmodynamics is a branch of knowldg that has its own postulats and tchniqus. W do not attmpt to giv hr vn an introduction to th fild. In this chaptr,

More information

VISUALIZATION OF DIFFERENTIAL GEOMETRY UDC 514.7(045) : : Eberhard Malkowsky 1, Vesna Veličković 2

VISUALIZATION OF DIFFERENTIAL GEOMETRY UDC 514.7(045) : : Eberhard Malkowsky 1, Vesna Veličković 2 FACTA UNIVERSITATIS Srs: Mchancs, Automatc Control Robotcs Vol.3, N o, 00, pp. 7-33 VISUALIZATION OF DIFFERENTIAL GEOMETRY UDC 54.7(045)54.75.6:59.688:59.673 Ebrhard Malkowsky, Vsna Vlčkovć Dpartmnt of

More information

Relate p and T at equilibrium between two phases. An open system where a new phase may form or a new component can be added

Relate p and T at equilibrium between two phases. An open system where a new phase may form or a new component can be added 4.3, 4.4 Phas Equlbrum Dtrmn th slops of th f lns Rlat p and at qulbrum btwn two phass ts consdr th Gbbs functon dg η + V Appls to a homognous systm An opn systm whr a nw phas may form or a nw componnt

More information

JEE-2017 : Advanced Paper 2 Answers and Explanations

JEE-2017 : Advanced Paper 2 Answers and Explanations DE 9 JEE-07 : Advancd Papr Answrs and Explanatons Physcs hmstry Mathmatcs 0 A, B, 9 A 8 B, 7 B 6 B, D B 0 D 9, D 8 D 7 A, B, D A 0 A,, D 9 8 * A A, B A B, D 0 B 9 A, D 5 D A, B A,B,,D A 50 A, 6 5 A D B

More information

COHORT MBA. Exponential function. MATH review (part2) by Lucian Mitroiu. The LOG and EXP functions. Properties: e e. lim.

COHORT MBA. Exponential function. MATH review (part2) by Lucian Mitroiu. The LOG and EXP functions. Properties: e e. lim. MTH rviw part b Lucian Mitroiu Th LOG and EXP functions Th ponntial function p : R, dfind as Proprtis: lim > lim p Eponntial function Y 8 6 - -8-6 - - X Th natural logarithm function ln in US- log: function

More information

Ερωτήσεις και ασκησεις Κεφ. 10 (για μόρια) ΠΑΡΑΔΟΣΗ 29/11/2016. (d)

Ερωτήσεις και ασκησεις Κεφ. 10 (για μόρια) ΠΑΡΑΔΟΣΗ 29/11/2016. (d) Ερωτήσεις και ασκησεις Κεφ 0 (για μόρια ΠΑΡΑΔΟΣΗ 9//06 Th coffcnt A of th van r Waals ntracton s: (a A r r / ( r r ( (c a a a a A r r / ( r r ( a a a a A r r / ( r r a a a a A r r / ( r r 4 a a a a 0 Th

More information

Radial Cataphoresis in Hg-Ar Fluorescent Lamp Discharges at High Power Density

Radial Cataphoresis in Hg-Ar Fluorescent Lamp Discharges at High Power Density [NWP.19] Radal Cataphorss n Hg-Ar Fluorscnt Lamp schargs at Hgh Powr nsty Y. Aura, G. A. Bonvallt, J. E. Lawlr Unv. of Wsconsn-Madson, Physcs pt. ABSTRACT Radal cataphorss s a procss n whch th lowr onzaton

More information

Stress-Based Finite Element Methods for Dynamics Analysis of Euler-Bernoulli Beams with Various Boundary Conditions

Stress-Based Finite Element Methods for Dynamics Analysis of Euler-Bernoulli Beams with Various Boundary Conditions 9 Strss-Basd Fnt Elmnt Mthods for Dynamcs Analyss of Eulr-Brnoull Bams wth Varous Boundary Condtons Abstract In ths rsarch, two strss-basd fnt lmnt mthods ncludng th curvatur-basd fnt lmnt mthod (CFE)

More information

First derivative analysis

First derivative analysis Robrto s Nots on Dirntial Calculus Chaptr 8: Graphical analysis Sction First drivativ analysis What you nd to know alrady: How to us drivativs to idntiy th critical valus o a unction and its trm points

More information

Physics 256: Lecture 2. Physics

Physics 256: Lecture 2. Physics Physcs 56: Lctur Intro to Quantum Physcs Agnda for Today Complx Numbrs Intrfrnc of lght Intrfrnc Two slt ntrfrnc Dffracton Sngl slt dffracton Physcs 01: Lctur 1, Pg 1 Constructv Intrfrnc Ths wll occur

More information

Electrochemical Equilibrium Electromotive Force. Relation between chemical and electric driving forces

Electrochemical Equilibrium Electromotive Force. Relation between chemical and electric driving forces C465/865, 26-3, Lctur 7, 2 th Sp., 26 lctrochmcal qulbrum lctromotv Forc Rlaton btwn chmcal and lctrc drvng forcs lctrochmcal systm at constant T and p: consdr G Consdr lctrochmcal racton (nvolvng transfr

More information

On the irreducibility of some polynomials in two variables

On the irreducibility of some polynomials in two variables ACTA ARITHMETICA LXXXII.3 (1997) On th irrducibility of som polynomials in two variabls by B. Brindza and Á. Pintér (Dbrcn) To th mmory of Paul Erdős Lt f(x) and g(y ) b polynomials with intgral cofficints

More information

A Distributed Real-time Database Index Algorithm Based on B+ Tree and Consistent Hashing

A Distributed Real-time Database Index Algorithm Based on B+ Tree and Consistent Hashing Avalabl onln at www.scncdrct.com Procda Engnrng 24 (2011) 171 176 2011 Intrnatonal Confrnc on Advancs n Engnrng A Dstrbutd Ral-tm Databas Indx Algorthm Basd on B+ Tr and Consstnt Hashng Xanhu L a, Cuhua

More information

Homotopy perturbation technique

Homotopy perturbation technique Comput. Mthods Appl. Mch. Engrg. 178 (1999) 257±262 www.lsvir.com/locat/cma Homotopy prturbation tchniqu Ji-Huan H 1 Shanghai Univrsity, Shanghai Institut of Applid Mathmatics and Mchanics, Shanghai 272,

More information

Study of Dynamic Aperture for PETRA III Ring K. Balewski, W. Brefeld, W. Decking, Y. Li DESY

Study of Dynamic Aperture for PETRA III Ring K. Balewski, W. Brefeld, W. Decking, Y. Li DESY Stud of Dnamc Aprtur for PETRA III Rng K. Balws, W. Brfld, W. Dcng, Y. L DESY FLS6 Hamburg PETRA III Yong-Jun L t al. Ovrvw Introducton Dnamcs of dampng wgglrs hoc of machn tuns, and optmzaton of stupol

More information

1973 AP Calculus AB: Section I

1973 AP Calculus AB: Section I 97 AP Calculus AB: Sction I 9 Minuts No Calculator Not: In this amination, ln dnots th natural logarithm of (that is, logarithm to th bas ).. ( ) d= + C 6 + C + C + C + C. If f ( ) = + + + and ( ), g=

More information

VLSI Implementation and Performance Evaluation of Low Pass Cascade & Linear Phase FIR Filter

VLSI Implementation and Performance Evaluation of Low Pass Cascade & Linear Phase FIR Filter Intrnatonal Journal of Engnrng and Tchncal Rsarch IJETR ISS: 3-869, Volum-3, Issu-6, Jun 5 VLSI Implmntaton and Prformanc Evaluaton of Low Pass Cascad & Lnar Phas Fltr Jaya Gupta, Arpan Shah, Ramsh Bhart

More information

Searching Linked Lists. Perfect Skip List. Building a Skip List. Skip List Analysis (1) Assume the list is sorted, but is stored in a linked list.

Searching Linked Lists. Perfect Skip List. Building a Skip List. Skip List Analysis (1) Assume the list is sorted, but is stored in a linked list. 3 3 4 8 6 3 3 4 8 6 3 3 4 8 6 () (d) 3 Sarching Linkd Lists Sarching Linkd Lists Sarching Linkd Lists ssum th list is sortd, but is stord in a linkd list. an w us binary sarch? omparisons? Work? What if

More information

An Efficient Algorithm for Mining Frequent Itemests over the Entire History of Data Streams

An Efficient Algorithm for Mining Frequent Itemests over the Entire History of Data Streams An Ecnt Algorthm or Mnng Frqunt Itmsts ovr th Entr Hstory o Data Strams Hua-Fu L, Suh-Yn L and Man-Kwan Shan Dpartmnt o Computr Scnc and Inormaton Engnrng, Natonal Chao-Tung Unvrsty, No. 00 Ta Hsuh Road,

More information

Problem Set 6 Solutions

Problem Set 6 Solutions 6.04/18.06J Mathmatics for Computr Scinc March 15, 005 Srini Dvadas and Eric Lhman Problm St 6 Solutions Du: Monday, March 8 at 9 PM in Room 3-044 Problm 1. Sammy th Shark is a financial srvic providr

More information

An Effective Technique for Enhancing Anti-Interference Performance of Adaptive Virtual Antenna Array

An Effective Technique for Enhancing Anti-Interference Performance of Adaptive Virtual Antenna Array 34 ACES JOURNAL VOL. 6 NO. 3 MARC 11 An Effctv Tchnqu for Enhancng Ant-Intrfrnc Prformanc of Adaptv Vrtual Antnna Array 1 Wnxng L 1 Ypng L 1 Ll Guo and Wnhua Yu 1 Collg of Informaton and Communcaton Engnrng

More information

are given in the table below. t (hours)

are given in the table below. t (hours) CALCULUS WORKSHEET ON INTEGRATION WITH DATA Work th following on notbook papr. Giv dcimal answrs corrct to thr dcimal placs.. A tank contains gallons of oil at tim t = hours. Oil is bing pumpd into th

More information

The Speed Bounds on Event Reporting in Mobile Sensor Networks with Energy Constraints

The Speed Bounds on Event Reporting in Mobile Sensor Networks with Energy Constraints Ths full txt papr was pr rvwd at th drcton of IEEE Communcatons Soct subjct mattr xprts for publcaton n th IEEE ICC 0 procdngs Th Spd Bounds on Evnt Rportng n Mobl Snsor Ntworks wth Enrg Constrants Y Xu

More information