An Integrated and Interactive Video Retrieval Framework with Hierarchical Learning Models and Semantic Clustering Strategy

Size: px
Start display at page:

Download "An Integrated and Interactive Video Retrieval Framework with Hierarchical Learning Models and Semantic Clustering Strategy"

Transcription

1 An Inegraed and Ineracve Vdeo Rereval Framewor wh Herarchcal Learnng Models and Semanc Cluserng Sraegy Na Zhao, Shu-Chng Chen, Me-Lng Shyu 2, Suar H. Rubn 3 Dsrbued Mulmeda Informaon Sysem Laboraory School of Compung and Informaon Scences Florda Inernaonal Unversy, Mam, FL 3399, USA 2 Deparmen of Elecrcal and Compuer Engneerng Unversy of Mam, Coral Gables, FL 3324, USA 3 Space and Naval Warfare Sysems Cener (SSC), San Dego, CA , USA {nzhao002, chens}@cs.fu.edu, 2 shyu@mam.edu, 3 suar.rubn@navy.ml Absrac In hs research, we propose an negraed and neracve framewor o manage and rereve large scale vdeo archves. The vdeo daa are modeled by a herarchcal learnng mechansm called HMMM (Herarchcal Marov Model Medaor) and ndexed by an nnovave semanc vdeo daabase cluserng sraegy. The cumulaed user feedbacs are reused o updae he affny relaonshps of he vdeo obecs as well as her nal sae probables. Correspondngly, boh he hgh level semancs and user percepons are employed n he vdeo cluserng sraegy. The clusered vdeo daabase s capable of provdng appealng mulmeda experence o he users because he modeled mulmeda daabase sysem can learn he user s preferences and neress neracvely.. Inroducon and Relaed Wor Wh he recen advances n mulmeda echnologes, he number of mulmeda fles and archves ncrease dramacally. Therefore, becomes an mporan research opc o mne and cluser he mulmeda daa, especally o accommodae he requremens of vdeo rereval n a dsrbued envronmen. Snce he mulmeda daabases may be dsrbued geographcally hrough he local newor or world-wde Inerne, he assocaed worloads could be que expensve when dealng wh complcaed vdeo queres. In parcular, semanc based vdeo rereval s mul-dscplnary and nvolves he negraon of vsual/audo feaures, emporal/spaal relaonshps, semanc evens/even paerns, hgh-level user percepons, ec. Therefore, s expeced o ulze a concepual daabase cluserng echnque o ndex and manage he mulmeda daabases such ha he relaed daa can be rereved ogeher and furhermore he communcaon coss n he query processng can be sgnfcanly reduced /06/$ IEEE. 438 Currenly, here exs approaches focusng on he cluserng echnques for he vdeo daa. For example, a herarchcal cluserng mehod for spors vdeo was presened []. Two levels of clusers are consruced where he op level s clusered by he color feaure and he boom level s clusered by he moon vecors. [2] descrbes a specral cluserng mehod o group vdeo shos no scenes based on her vsual smlary and emporal relaonshps. In [5], he algorhms are proposed for unsupervsed dscovery of he vdeo srucure by modelng he evens and her sochasc srucures n vdeo sequences va usng Herarchcal Hdden Marov Models (HHMM). Based on our bes nowledge, mos of he exsng researches produce he clusers manly on low-level and/or md-level feaures, and do no consder hgh-level conceps or user percepons n he cluserng procedure. Ths brngs he problem of semanc gap. Relevance feedbac s an effecve mehod o narrow down hs semanc gap. However, mos of he exsng relevance feedbac sysems are only capable of provdng real-me updaes on he rereval resuls whou any furher mprovemen of he overall sysem performance. In addon, mulmeda daabases may no be effcenly modeled n hese approaches even afer he cluserng echnque s appled. In hs paper, an negraed and neracve vdeo rereval framewor s proposed o effcenly organze, model, and rereve he conen of a large scale mulmeda daabase. The core of our proposed framewor s a learnng mechansm called HMMM (Herarchcal Marov Model Medaor) [6] and an nnovave vdeo cluserng sraegy. HMMM models he vdeo daabase; whle he cluserng sraegy groups vdeo daa wh smlar characerscs no clusers ha exhb ceran hgh level semancs. The HMMM srucure s hen exended by addng an addonal level o represen he clusers and her relaonshps. The proposed framewor s desgned o accommodae advanced queres va consderng he hgh level semanc meanng. Frs of all, s capable of searchng semanc

2 evens or even paerns consderng her populary by evaluang her access frequences n he large amoun of hsorcal queres. Second, he users can choose one or more example paerns wh her ancpaed feaures from he nal rereved resuls, and hen ssue he nex round of query. I can search and re-ran he canddae paerns whch nvolve he smlar aspecs wh he posve examples reflecng he user s neress. Thrd, vdeo cluserng can be conduced o furher reduce he searchng me especally when dealng wh he op- smlary rerevals. As he HMMM mechansm helps o raverse he mos opmzed pah o perform he rereval, he proposed framewor can only search several clusers for he canddae resuls whou raversng all he pahs o chec he whole daabase. Ths paper s organzed as follows: Secon 2 presens he overall framewor of our proposed research. In Secon 3, he dealed echnques are furher expanded by nroducng HMMM model and explanng he cluserng sraegy. Moreover, he rereval algorhm and example are also ncluded. Secon 4 analyzes he expermenal resuls. Fnally, conclusons are summarzed n Secon Overall Framewor Fgure demonsraes he overall worflow of he proposed framewor. In hs framewor, he soccer vdeos are frs segmened no dsnc vdeo shos and her lowlevel vdeo/audo feaures are exraced. A mulmeda daa mnng approach s ulzed o pre-process he vdeo shos o ge an nal canddae pool for he poenal mporan evens. Afer ha, a se of nal even labels wll be gven o some of he shos, where no all of hese labels are correc. All of hese daa and nformaon wll be fed no hs framewor for even paern searchng and vdeo rereval purposes. The vdeos ncluded n he canddae pool are modeled n he s level of MMM (Marov Model Medaor) models, whereas he vdeos are modeled n he 2 nd level. Afer nalzng he s level and 2 nd level of MMM models, he users are allowed o ssue he even or even paern queres. Furhermore, he users can selec her neresed even paerns n he nal resuls and re-ssue he query o refne he rereval resuls and her ranngs. Ths sep s also recognzed as onlne learnng. These user seleced sho sequences are sored as posve paerns for he fuure offlne ranng. Afer a ceran amoun of queres and feedbacs, he proposed framewor s able o perform he offlne ranng. The hsorcal queres and user access records are ulzed o updae he affny relaonshps of he vdeos/vdeo shos as well as her nal sae probables. Thereafer, boh he semanc evens and he hgh level user percepons are employed o consruc he vdeo clusers, whch are hen modeled by a hgher level (3 rd level) of he MMM model. In he meanwhle, he 2 nd level MMM model are dvded no a se of sub-models based on he clusered vdeo groups. The clusered daabase and he updaed HMMM mechansm are capable of provdng appealng mulmeda experence o he users because he modeled mulmeda daabase sysem learns he user s preferences and neress neracvely va reusng he hsorcal queres. Vdeo Sho Segmenaon Sho Feaure Exracon Daa Cleanng Even Deecon Even/Paern Queres Feedbac Accumulae Offlne Tranng Vdeo Cluserng 3. Vdeo Daabase Cluserng 3.. Vdeo Daabase Modelng Inalze he s level and 2 nd level MMM models Updae he s level and 2 nd level MMM models Consruc he 3 rd level MMM models and Updae he 2 nd level MMM models Provde enhanced performance for he Top- even paern rereval Fgure. Overall worflow for he proposed approach In our prevous sudes, we have successfully appled he MMM (Marov Model Medaor) model n mage daabase cluserng [4], and expanded MMM o HMMM [6] o model a vdeo daabase whch s formalzed as an 8- uple λ = ( d, S, F, A, B, Π, O, L). Le n denoe he level number, where n d and d s he number of levels n an HMMM. The n h level of HMMM may conan one or more MMM models o represen he ses of dsnc mulmeda obecs and her assocae feaures. In [6], d s se as 2. S n he lowes-level MMM represens he se of vdeo shos, and he feaure se F consss of vsual/audo feaures. Whle n he 2 nd level MMM, S 2 descrbes he se of vdeos n he daabase, and F 2 conans he semanc evens deeced n he vdeo collecon. Each of he MMM models ncorporaes a se of marces for affny relaonshps (A n ), feaure values (B n ), and nal sae probably dsrbuons (Π n ). In addon, O and L are desgned for he relaonshp descrpon beween wo adacen levels. O (O,2 ) ncludes he weghs of mporance for he low-level feaures (F ) when descrbng he hgh-level semanc evens (F 2 ). L (L,2 ) descrbes he ln condons beween he vdeos (S 2 ) and he vdeo shos (S ). HMMM model carres ou a sochasc and 439

3 dynamc process n boh search and smlary calculaon, where always res o raverse he pah wh he larges possbly. Therefore, can asss n rerevng more accurae paerns qucly wh lower compuaonal coss. The specfc desgn of HMMM helps no only he general even queres, bu also he rereval of emporal based semanc even paerns. Anoher sgnfcan advanage of HMMM s s neracve feedbac and learnng sraeges, whch can profcenly assure he connuous mprovemens of he overall performance. The users are capable of provdng her own feedbacs such ha he sysem can be raned eher onlne or offlne. The onlne learnng mechansm creaes an ndvdual MMM nsance by usng he vdeo shos ha a user prefers. All of he users feedbacs are effcenly accumulaed and ready for he offlne sysem ranng process. In hs research, he large amoun of user feedbacs wll be reused n vdeo daabase cluserng o furher mprove he overall rereval performance and reduce he searchng space and me Concepual Vdeo Cluserng Smlary Measuremen In hs proposed framewor, a vdeo s reaed as an ndvdual daabase n a dsrbued mulmeda daabase sysem, where s vdeo shos are he daa nsances n he daabase. Accordngly, a smlary measure beween wo vdeos s defned as a value ndcang he leness of hese wo vdeos wh respec o her concepual conens. I s calculaed by evaluang her posve evens and even paerns n he hsorcal queres. Tha s, f wo vdeos conss of he same even(s) and/or even paern(s) and are accessed ogeher frequenly, s consdered ha hese wo vdeos are closely relaed and her smlary score should be hgh. Assume here are H user queres ssued hrough he vdeo rereval framewor, where he se of all he query paerns s denoed as QS. In order o refne her rereved resuls n real-me, he users mar her preferred even paerns as posve before mang he nex query. By evaluang he ssued query ses and her assocaed posve paerns, he smlary measure s defned as follows. Le v and v be wo vdeos, and X={x,, x m } and Y={y,, y n } be he ses of vdeo shos belongng o v and v ( X v, Y v ), where m and n are he numbers of annoaed vdeo shos n v and v. Denoe a query wh an observaon sequence (semanc even paern) wh C semanc evens as Q = e, e,..., e }, where Q QS. Le R be he se of { 2 C G posve paerns ha a user seleced from he nal rereval resuls for query Q. Ths can be represened by a marx of sze G C, G, C. As shown n Equaon (), each row of R represens an even sho sequence ha he user mared as posve, and each column ncludes he canddae even shos whch correspond o he requesed even n he query paern. { s, s2,..., sc } { s, s2,..., sc } R =. ()... G G G { s, s2,..., sc } Based on he above assumpons, he vdeo smlary funcon s defned as below. Defnon : SV(v, v ), he smlary measure beween wo vdeos, s defned by evaluang he probables of fndng he same even paern Q from v and v n he same query for all he query paerns n QS. ( v ) P( Q v ) FA( ) SV( v, v ) = P Q H, (2) Q QS where H, and FA (H ) s an adusng facor. P ( Q ) v and ( Q ) v P represen he occurrence probables of fndng Q from v and v, where he occurrence probably can be obaned by summng he on probables over all he possble saes [3]. In order o calculae hs value, we need o selec all he subses wh C even shos from he posve paern se R, whch also belong o v or v. Tha s, X = { x, x,..., x } and Y = { y, y,..., y }, where X X, 2 C 2 C, Y Y, X R Y R. If hese paerns do no exs, hen he probably value s se as 0 auomacally. ( Q v ) = P( Q, X v ) = P( Q X, v ) P( X v ) P. (3) all X all X Assume he sascal ndependence of he observaons, and gven he sae sequence of X = { x, x2,..., xc }, Equaon (4) gves he probably of X gven v. P C C ( X v ) = P( x x ) P( x ) = A ( x, x ) π ( ) + x = = Here, ( ) +. (4) P x x + represens he probably of rerevng a vdeo sho x gven ha he curren vdeo sho s x. I + corresponds o he A ( x, x + ) enry n he relaonshp P s he nal probably for vdeo sho x, marx. ( x ).e., π ( ). Equaon (5) gves he probably of an x observaon sequence (semanc even paern) C ( X, v ) = P Q = Q. P( e x ), (5) 440

4 where P e x ) ndcaes he probably of observng a ( semanc even e from a vdeo sho x. Ths value s compued usng a smlary measure by consderng lowlevel and md-level feaures. However, n hs approach, snce he users already mared hese vdeo shos as he evens hey requesed and preferred, he probables of observng he semanc evens are smply se o Cluserng Sraegy Consderng a large scale vdeo daabase, s a sgnfcan ssue o cluser smlar vdeos ogeher o speed up he smlary search. As we saed before, a wo-level HMMM has been consruced o model vdeo and vdeo shos. Furhermore, a vdeo daabase cluserng sraegy whch s raversal-based and greedy s proposed. Y p=m? N Inpus: () D: daabase ncludng M vdeos (2) Z: Maxmum sze of a cluser (Z 2) p=0; n=0; n++; sar a new cluser: CC n {}; CC n CC n {v }; q=; p++; where v has he larges Π 2(v ) value n D D D-{v }; Add v o CC n: CC n CC n {v } ; where v has he larges value of A ( v, v ) SV ( v, v ) n D 2 v v ; D D-{v }; p++; q++; Oupus: () n: Number of vdeo clusers (2) CC ( n): vdeo clusers q=z? p=m? Fgure 2. The proposed concepual vdeo daabase cluserng procedure As llusraed n Fgure 2, he proposed vdeo daabase cluserng echnque conans he followng seps. Gven he vdeo daabase D wh M vdeos and he maxmum sze of he vdeo daabase cluser as Z (Z 2), he mechansm: a) Inalze he parameers as p=0; n=0, where p denoes he number of vdeos beng clusered, and n represens he cluser number. b) Se n=n+. Search he curren vdeo daabase D for he vdeo v wh he larges saonary probably N Y N Y Π 2 (v ), and hen sars a new cluser CC n wh hs vdeo (CC n = { }; CC n CC n {v }). Inalze he parameer as q=, where q represens he number of vdeos n he curren cluser. c) Remove v from daabase D (D D-{v }). Chec f p=m. If yes, oupu he clusers. If no, go o sep d). d) Search for v, whch has he larges A ( v, v ) SV ( v, v ) n D. Add v o he curren 2 cluser CC n (CC n CC n {v }). e) v v, where v represens he mos recen clusered vdeo. Every me when a vdeo s assgned o a cluser, s auomacally removed from D (D D {v }). f) p++ and q++. Chec f p=m. If yes, oupu he cluserng resuls. If no, chec f q=z. If yes, goes o sep b) o sar a new cluser. If no, goes o sep d) o add anoher vdeo n he curren cluser. g) If here s no un-clusered vdeo lef n he curren daabase, oupu he curren clusers Ineracve Rereval upon Vdeo Clusers Iner-Cluser Relaonshps In hs research, he HMMM model s exended by he 3 rd level MMM o mprove he overall rereval performance. In he 3 rd level MMM (d = 3), he saes (S 3 ) denoes he vdeo clusers. Marx A 3 descrbes he relaonshps beween each par of clusers. Defnon 2: Assume CC m and CC n are wo vdeo clusers n he vdeo daabase D. Ther relaonshp s denoed as an enry n he affny marx A 3, whch can be compued by Equaons (6) and (7). Here, SC s he funcon ha calculaes he smlary score beween wo vdeo clusers. SC( CCm, CCn) = ( Π2 ( v ) max ( A2 ( v, v ) SV ( v, v )))/ M, S F A B Π v CC m where CC m v CCn D, CC D. (6) SC( CCm, CCn ) A3 ( CCm, CCn ) =. (7) SC( CC, CC ) CC D Table. 3-Level HMMM Model s Level MMM 2 nd Level MMM 3 rd Level MMM Sae se of vdeo shos Low level vsual/audo feaures Temporal based sae ranson probably beween vdeo shos Formalzed feaure values Inal sae probably dsrbuon for vdeo shos n m Sae se of Vdeos Semanc evens (conceps) Affny relaonshp beween vdeos Annoaed even numbers Inal sae probably dsrbuon for vdeos Sae se of vdeo clusers - Affny relaonshp beween vdeo clusers - Inal sae probably dsrbuon for vdeo clusers 44

5 The marx Π 3 can be consruced o represen he nal sae probably of he clusers. The calculaon of Π 3 s smlar o he ones for Π and Π 2. In addon, marx L 2,3 can also be consruced o llusrae he ln condons beween he 2 nd level MMMs and he 3 rd level MMM. As demonsraed n Table, he MMM models n dfferen levels of he 3-level HMMM descrbe dsnc obecs and represen dfferen meanngs Rereval hrough Clusered Vdeo Daabase Gven an example sho sequence Q = { s, s,..., s } 2 C whch represens he even paern as e, e,..., e } such { 2 C ha s descrbes e ( C), and hey follow he emporal sequence as T T... T. Assume ha a s s 2 s C user wans o fnd op- relaed sho sequences whch follow he smlar paerns. In our proposed rereval algorhm, a recursve process s conduced o raverse he HMMM daabase model and fnd he op canddae resuls. As shown n Fgure 3, a lace based srucure for he overall vdeo daabase can be consruced. Assume he ransons are sored based on her edge weghs [6], and he rereval algorhm wll raverse he edge wh a hgher wegh each me. For example, n Fgure 4, we assume ha he edge weghs sasfy w(s, s 2 ) w(s, s 4 ) w(s, s 7 ). The rereval algorhm can be descrbed as below.. Search for he frs canddae cluser, frs canddae vdeo and frs canddae vdeo sho by checng marces Π 3, Π 2, B 2, Π and B. 2. If he paern s no complee, connue search for he nex even (vdeo sho) va compung he edge weghs by checng A. 3. If he canddae paern has been compleed, he sysem goes bac sae by sae and checs for oher possble pahs. The sysem also checs f here are already canddae paerns beng rereved. If yes, he sysem sops searchng and goes o Sep If here s no more possbles n he curren vdeo, hen mar hs vdeo wh a searched flag and chec A 2 and B 2 o fnd nex canddae vdeo. 5. If all he vdeos are searched n he curren cluser, hen mar he curren cluser as searched cluser and chec A 3 o fnd he nex canddae vdeo cluser. 6. Once paerns are rereved, or here are no more possbles n he daabase, he sysem rans he canddae paerns va calculang he smlary scores [6] and oupus he canddae paerns. As shown n he Fgure 4, he yellow cells nclude he pahs he algorhm raversed. Furhermore, we desgned a funcon o fll n he mssed cells by copyng he corresponden shos n he prevous canddae paerns. Fnally, 6 complee canddae paerns are generaed. Once canddae paerns are generaed, he sysem does no need o raverse any oher clusers or vdeos. Therefore, sgnfcanly reduces he searchng spaces and acceleraes he searchng speed. CC CC 2 Cluser Vdeo T : Even T 2: Even 2 v v 2 v 3 Canddae vdeo cluser Canddae vdeo sae Vdeo sho sae whch maches he expeced even Transon whch maches he expeced emporal even Transon whch goes o search he nex canddae vdeo Fgure 3. Lace srucure of he clusered vdeo daabase Fgure 4. Resul paerns and he raverse pah s s 9 s 0 4. EXPERIMENTAL RESULTS T 3: Even 3 We have bul up a soccer vdeo daabase wh oally 45 vdeos, whch conans 8977 vdeo shos. A rereval sysem has also been mplemened for he sysem ranng and expermenal ess. Toally 50 ses of hsorcal queres were ssued and user feedbacs were reurned wh her preferred paerns, whch cover all of he 45 vdeos and 259 dsnc vdeo shos. In he cluserng process, we defne he cluser sze as 0 and he expeced resul paern number as =60. As shown n Fgure 5, we use leers G, F, and C o represen Goal, Free c, Corner c evens, respecvely. Therefore, he x-axs represens dfferen query paerns, e.g., G means a query o search for Goal evens; FG means a query o search for he even paern where a Free c followed by a Goal ; and CGF means a query paern of a Corner c even, followed by a Goal and hen a Free c, ec. For each query paern, we ssued 0 queres o compue he average execuon me n mllseconds. As llusraed n Fgure 5, he query paerns wh fewer even numbers wll be execued n less me as expeced. In addon, he execuon me of he sysem wh clusers s less han ha of he sysem whou clusers, ndcang ha our proposed approach effecvely s 2 s 4 s 7 s 3 s 5 s 6 s 8 s s 2 s 3 R Cluser Vdeo Even Even 2 Even 3 CC v s s 2 s 3 2 CC v s s 4 s 5 3 CC v s s 4 s 6 4 CC v s s 7 s 8 5 CC v s 9 s 7 s 8 6 CC 2 v 3 s s 2 s 3 442

6 groups relevan vdeos n he vdeo clusers so ha only he relevan clusers and her member vdeos wll need o be searched. Therefore, he searchng space s dramacally decreased, and he execuon of he queres becomes faser. 200 Execuon Tme Comparson Average Execuon Tme (mllseconds) whou Clusers wh Clusers (a) Query over non-clusered soccer vdeo daabase 0 G C F FG CG GC CGF FCG FGC Query Paern Fgure 5. Comparson of he average execuon me For he query paern ( Corner c followed by a Goal ), Fgure 6(a) demonsraes he frs screen of rereval resuls over he non-clusered soccer vdeo daabase; whle Fgure 6(b) shows he query resuls over he clusered daabase. I can be clearly seen ha he query resuls n he same cluser represen he smlar vsual clues, whch are mned from he hsorcal queres and feedbacs, and accordngly represen user preferences. 5. CONCLUSIONS In hs paper, an neracve vdeo rereval sysem s proposed whch ncorporaes he concepual vdeo cluserng sraegy and he HMMM herarchcal learnng mechansm. Ths proposed framewor s able o reuse he cumulaed user feedbacs o cluser he vdeos, such ha he overall sysem no only learns he user percepons, bu also ges a good daabase srucure va adopng he cluserng echnque. The HMMM-based daabase model s consruced o suppor he concepual vdeo daabase cluserng. In he meanwhle, he cluserng echnque helps o furher mprove he daabase srucure va addng a new level o model he vdeo clusers. The expermens show ha our proposed approach helps accelerae he rereval speed wh provdng decen rereval resuls. 6. ACKOWLEDGEMENTS For Shu-Chng Chen, hs research was suppored n par by NSF EIA and HRD For Me-Lng Shyu, hs research was suppored n par by NSF ITR (Medum) IIS For Suar Rubn, hs research was suppored n par by an ONR ILIR gran. (b) Query over clusered soccer vdeo daabase Fgure 6. Soccer vdeo rereval sysem nerfaces 7. REFERENCES [] C.-W. Ngo, T.-C. Pong and H.-J. Zhang, On Cluserng and Rereval of Vdeo Shos, In Proc. of he 9 h ACM Inernaonal Conference on Mulmeda, Oawa, Canada, 200, pp [2] J.-M. Odobez, D. Gaca-Perez, and M. Gullemo, Vdeo Sho Cluserng usng Specral Mehods, In Proc. of 3rd Inernaonal Worshop on Conen-Based Mulmeda Indexng (CBMI), Rennes, France, 2003, pp [3] L. Rabner and B.-H. Juang. Fundamenals of Speech Recognon, Prence Hall, 993, ISBN: [4] M.-L. Shyu, S.-C. Chen, M. Chen, and C. Zhang, Affny Relaon Dscovery n Image Daabase Cluserng and Conenbased Rereval, In Proc. of ACM Mulmeda 2004 Conference, New Yor, USA, pp [5] L. Xe, S.-F. Chang, A. Dvaaran, and H. Sun, Unsupervsed Dscovery of Mullevel Sascal Vdeo Srucures Usng Herarchcal Hdden Marov Models, In Proc. of IEEE Inernaonal Conference on Mulmeda and Expo (ICME), vol. 3, pp , July [6] N. Zhao, S.-C. Chen and M.-L. Shyu, Vdeo Daabase Modelng and Temporal Paern Rereval Usng Herarchcal Marov Model Medaor, In Proc. of he Frs IEEE Inernaonal Worshop on Mulmeda Daabases and Daa Managemen (MDDM), n conuncon wh IEEE Inernaonal Conference on Daa Engneerng (ICDE), Alana, USA,

Computing Relevance, Similarity: The Vector Space Model

Computing Relevance, Similarity: The Vector Space Model Compung Relevance, Smlary: The Vecor Space Model Based on Larson and Hears s sldes a UC-Bereley hp://.sms.bereley.edu/courses/s0/f00/ aabase Managemen Sysems, R. Ramarshnan ocumen Vecors v ocumens are

More information

Clustering (Bishop ch 9)

Clustering (Bishop ch 9) Cluserng (Bshop ch 9) Reference: Daa Mnng by Margare Dunham (a slde source) 1 Cluserng Cluserng s unsupervsed learnng, here are no class labels Wan o fnd groups of smlar nsances Ofen use a dsance measure

More information

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore.

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore. Ths documen s downloaded from DR-NTU, Nanyang Technologcal Unversy Lbrary, Sngapore. Tle A smplfed verb machng algorhm for word paron n vsual speech processng( Acceped verson ) Auhor(s) Foo, Say We; Yong,

More information

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon

More information

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model Probablsc Model for Tme-seres Daa: Hdden Markov Model Hrosh Mamsuka Bonformacs Cener Kyoo Unversy Oulne Three Problems for probablsc models n machne learnng. Compung lkelhood 2. Learnng 3. Parsng (predcon

More information

Introduction ( Week 1-2) Course introduction A brief introduction to molecular biology A brief introduction to sequence comparison Part I: Algorithms

Introduction ( Week 1-2) Course introduction A brief introduction to molecular biology A brief introduction to sequence comparison Part I: Algorithms Course organzaon Inroducon Wee -2) Course nroducon A bref nroducon o molecular bology A bref nroducon o sequence comparson Par I: Algorhms for Sequence Analyss Wee 3-8) Chaper -3, Models and heores» Probably

More information

Variants of Pegasos. December 11, 2009

Variants of Pegasos. December 11, 2009 Inroducon Varans of Pegasos SooWoong Ryu bshboy@sanford.edu December, 009 Youngsoo Cho yc344@sanford.edu Developng a new SVM algorhm s ongong research opc. Among many exng SVM algorhms, we wll focus on

More information

Solution in semi infinite diffusion couples (error function analysis)

Solution in semi infinite diffusion couples (error function analysis) Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of

More information

Attribute Reduction Algorithm Based on Discernibility Matrix with Algebraic Method GAO Jing1,a, Ma Hui1, Han Zhidong2,b

Attribute Reduction Algorithm Based on Discernibility Matrix with Algebraic Method GAO Jing1,a, Ma Hui1, Han Zhidong2,b Inernaonal Indusral Informacs and Compuer Engneerng Conference (IIICEC 05) Arbue educon Algorhm Based on Dscernbly Marx wh Algebrac Mehod GAO Jng,a, Ma Hu, Han Zhdong,b Informaon School, Capal Unversy

More information

Robustness Experiments with Two Variance Components

Robustness Experiments with Two Variance Components Naonal Insue of Sandards and Technology (NIST) Informaon Technology Laboraory (ITL) Sascal Engneerng Dvson (SED) Robusness Expermens wh Two Varance Componens by Ana Ivelsse Avlés avles@ns.gov Conference

More information

Cubic Bezier Homotopy Function for Solving Exponential Equations

Cubic Bezier Homotopy Function for Solving Exponential Equations Penerb Journal of Advanced Research n Compung and Applcaons ISSN (onlne: 46-97 Vol. 4, No.. Pages -8, 6 omoopy Funcon for Solvng Eponenal Equaons S. S. Raml *,,. Mohamad Nor,a, N. S. Saharzan,b and M.

More information

Hidden Markov Models

Hidden Markov Models 11-755 Machne Learnng for Sgnal Processng Hdden Markov Models Class 15. 12 Oc 2010 1 Admnsrva HW2 due Tuesday Is everyone on he projecs page? Where are your projec proposals? 2 Recap: Wha s an HMM Probablsc

More information

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy

More information

Learning Objectives. Self Organization Map. Hamming Distance(1/5) Introduction. Hamming Distance(3/5) Hamming Distance(2/5) 15/04/2015

Learning Objectives. Self Organization Map. Hamming Distance(1/5) Introduction. Hamming Distance(3/5) Hamming Distance(2/5) 15/04/2015 /4/ Learnng Objecves Self Organzaon Map Learnng whou Exaples. Inroducon. MAXNET 3. Cluserng 4. Feaure Map. Self-organzng Feaure Map 6. Concluson 38 Inroducon. Learnng whou exaples. Daa are npu o he syse

More information

Advanced Machine Learning & Perception

Advanced Machine Learning & Perception Advanced Machne Learnng & Percepon Insrucor: Tony Jebara SVM Feaure & Kernel Selecon SVM Eensons Feaure Selecon (Flerng and Wrappng) SVM Feaure Selecon SVM Kernel Selecon SVM Eensons Classfcaon Feaure/Kernel

More information

On One Analytic Method of. Constructing Program Controls

On One Analytic Method of. Constructing Program Controls Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna

More information

Math 128b Project. Jude Yuen

Math 128b Project. Jude Yuen Mah 8b Proec Jude Yuen . Inroducon Le { Z } be a sequence of observed ndependen vecor varables. If he elemens of Z have a on normal dsrbuon hen { Z } has a mean vecor Z and a varancecovarance marx z. Geomercally

More information

Using Fuzzy Pattern Recognition to Detect Unknown Malicious Executables Code

Using Fuzzy Pattern Recognition to Detect Unknown Malicious Executables Code Usng Fuzzy Paern Recognon o Deec Unknown Malcous Execuables Code Boyun Zhang,, Janpng Yn, and Jngbo Hao School of Compuer Scence, Naonal Unversy of Defense Technology, Changsha 40073, Chna hnxzby@yahoo.com.cn

More information

Fall 2010 Graduate Course on Dynamic Learning

Fall 2010 Graduate Course on Dynamic Learning Fall 200 Graduae Course on Dynamc Learnng Chaper 4: Parcle Flers Sepember 27, 200 Byoung-Tak Zhang School of Compuer Scence and Engneerng & Cognve Scence and Bran Scence Programs Seoul aonal Unversy hp://b.snu.ac.kr/~bzhang/

More information

Robust and Accurate Cancer Classification with Gene Expression Profiling

Robust and Accurate Cancer Classification with Gene Expression Profiling Robus and Accurae Cancer Classfcaon wh Gene Expresson Proflng (Compuaonal ysems Bology, 2005) Auhor: Hafeng L, Keshu Zhang, ao Jang Oulne Background LDA (lnear dscrmnan analyss) and small sample sze problem

More information

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he

More information

An introduction to Support Vector Machine

An introduction to Support Vector Machine An nroducon o Suppor Vecor Machne 報告者 : 黃立德 References: Smon Haykn, "Neural Neworks: a comprehensve foundaon, second edon, 999, Chaper 2,6 Nello Chrsann, John Shawe-Tayer, An Inroducon o Suppor Vecor Machnes,

More information

An Effective TCM-KNN Scheme for High-Speed Network Anomaly Detection

An Effective TCM-KNN Scheme for High-Speed Network Anomaly Detection Vol. 24, November,, 200 An Effecve TCM-KNN Scheme for Hgh-Speed Nework Anomaly eecon Yang L Chnese Academy of Scences, Bejng Chna, 00080 lyang@sofware.c.ac.cn Absrac. Nework anomaly deecon has been a ho

More information

TSS = SST + SSE An orthogonal partition of the total SS

TSS = SST + SSE An orthogonal partition of the total SS ANOVA: Topc 4. Orhogonal conrass [ST&D p. 183] H 0 : µ 1 = µ =... = µ H 1 : The mean of a leas one reamen group s dfferen To es hs hypohess, a basc ANOVA allocaes he varaon among reamen means (SST) equally

More information

Bernoulli process with 282 ky periodicity is detected in the R-N reversals of the earth s magnetic field

Bernoulli process with 282 ky periodicity is detected in the R-N reversals of the earth s magnetic field Submed o: Suden Essay Awards n Magnecs Bernoull process wh 8 ky perodcy s deeced n he R-N reversals of he earh s magnec feld Jozsef Gara Deparmen of Earh Scences Florda Inernaonal Unversy Unversy Park,

More information

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005 Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc

More information

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Journal of Appled Mahemacs and Compuaonal Mechancs 3, (), 45-5 HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Sansław Kukla, Urszula Sedlecka Insue of Mahemacs,

More information

Appendix to Online Clustering with Experts

Appendix to Online Clustering with Experts A Appendx o Onlne Cluserng wh Expers Furher dscusson of expermens. Here we furher dscuss expermenal resuls repored n he paper. Ineresngly, we observe ha OCE (and n parcular Learn- ) racks he bes exper

More information

Linear Response Theory: The connection between QFT and experiments

Linear Response Theory: The connection between QFT and experiments Phys540.nb 39 3 Lnear Response Theory: The connecon beween QFT and expermens 3.1. Basc conceps and deas Q: ow do we measure he conducvy of a meal? A: we frs nroduce a weak elecrc feld E, and hen measure

More information

Testing a new idea to solve the P = NP problem with mathematical induction

Testing a new idea to solve the P = NP problem with mathematical induction Tesng a new dea o solve he P = NP problem wh mahemacal nducon Bacground P and NP are wo classes (ses) of languages n Compuer Scence An open problem s wheher P = NP Ths paper ess a new dea o compare he

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4 CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped

More information

WiH Wei He

WiH Wei He Sysem Idenfcaon of onlnear Sae-Space Space Baery odels WH We He wehe@calce.umd.edu Advsor: Dr. Chaochao Chen Deparmen of echancal Engneerng Unversy of aryland, College Par 1 Unversy of aryland Bacground

More information

Reactive Methods to Solve the Berth AllocationProblem with Stochastic Arrival and Handling Times

Reactive Methods to Solve the Berth AllocationProblem with Stochastic Arrival and Handling Times Reacve Mehods o Solve he Berh AllocaonProblem wh Sochasc Arrval and Handlng Tmes Nsh Umang* Mchel Berlare* * TRANSP-OR, Ecole Polyechnque Fédérale de Lausanne Frs Workshop on Large Scale Opmzaon November

More information

FI 3103 Quantum Physics

FI 3103 Quantum Physics /9/4 FI 33 Quanum Physcs Aleander A. Iskandar Physcs of Magnesm and Phooncs Research Grou Insu Teknolog Bandung Basc Conces n Quanum Physcs Probably and Eecaon Value Hesenberg Uncerany Prncle Wave Funcon

More information

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes.

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes. umercal negraon of he dffuson equaon (I) Fne dfference mehod. Spaal screaon. Inernal nodes. R L V For hermal conducon le s dscree he spaal doman no small fne spans, =,,: Balance of parcles for an nernal

More information

Relative controllability of nonlinear systems with delays in control

Relative controllability of nonlinear systems with delays in control Relave conrollably o nonlnear sysems wh delays n conrol Jerzy Klamka Insue o Conrol Engneerng, Slesan Techncal Unversy, 44- Glwce, Poland. phone/ax : 48 32 37227, {jklamka}@a.polsl.glwce.pl Keywor: Conrollably.

More information

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!") i+1,q - [(!

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!) i+1,q - [(! ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL The frs hng o es n wo-way ANOVA: Is here neracon? "No neracon" means: The man effecs model would f. Ths n urn means: In he neracon plo (wh A on he horzonal

More information

Lecture 2 L n i e n a e r a M od o e d l e s

Lecture 2 L n i e n a e r a M od o e d l e s Lecure Lnear Models Las lecure You have learned abou ha s machne learnng Supervsed learnng Unsupervsed learnng Renforcemen learnng You have seen an eample learnng problem and he general process ha one

More information

Volatility Interpolation

Volatility Interpolation Volaly Inerpolaon Prelmnary Verson March 00 Jesper Andreasen and Bran Huge Danse Mares, Copenhagen wan.daddy@danseban.com brno@danseban.com Elecronc copy avalable a: hp://ssrn.com/absrac=69497 Inro Local

More information

Chapter 6: AC Circuits

Chapter 6: AC Circuits Chaper 6: AC Crcus Chaper 6: Oulne Phasors and he AC Seady Sae AC Crcus A sable, lnear crcu operang n he seady sae wh snusodal excaon (.e., snusodal seady sae. Complee response forced response naural response.

More information

Hidden Markov Models Following a lecture by Andrew W. Moore Carnegie Mellon University

Hidden Markov Models Following a lecture by Andrew W. Moore Carnegie Mellon University Hdden Markov Models Followng a lecure by Andrew W. Moore Carnege Mellon Unversy www.cs.cmu.edu/~awm/uorals A Markov Sysem Has N saes, called s, s 2.. s N s 2 There are dscree meseps, 0,, s s 3 N 3 0 Hdden

More information

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim Korean J. Mah. 19 (2011), No. 3, pp. 263 272 GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS Youngwoo Ahn and Kae Km Absrac. In he paper [1], an explc correspondence beween ceran

More information

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations.

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations. Soluons o Ordnary Derenal Equaons An ordnary derenal equaon has only one ndependen varable. A sysem o ordnary derenal equaons consss o several derenal equaons each wh he same ndependen varable. An eample

More information

Theoretical Analysis of Biogeography Based Optimization Aijun ZHU1,2,3 a, Cong HU1,3, Chuanpei XU1,3, Zhi Li1,3

Theoretical Analysis of Biogeography Based Optimization Aijun ZHU1,2,3 a, Cong HU1,3, Chuanpei XU1,3, Zhi Li1,3 6h Inernaonal Conference on Machnery, Maerals, Envronmen, Boechnology and Compuer (MMEBC 6) Theorecal Analyss of Bogeography Based Opmzaon Aun ZU,,3 a, Cong U,3, Chuanpe XU,3, Zh L,3 School of Elecronc

More information

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov June 7 e-ournal Relably: Theory& Applcaons No (Vol. CONFIDENCE INTERVALS ASSOCIATED WITH PERFORMANCE ANALYSIS OF SYMMETRIC LARGE CLOSED CLIENT/SERVER COMPUTER NETWORKS Absrac Vyacheslav Abramov School

More information

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL Sco Wsdom, John Hershey 2, Jonahan Le Roux 2, and Shnj Waanabe 2 Deparmen o Elecrcal Engneerng, Unversy o Washngon, Seale, WA, USA

More information

Lecture Slides for INTRODUCTION TO. Machine Learning. ETHEM ALPAYDIN The MIT Press,

Lecture Slides for INTRODUCTION TO. Machine Learning. ETHEM ALPAYDIN The MIT Press, Lecure Sldes for INTRDUCTIN T Machne Learnng ETHEM ALAYDIN The MIT ress, 2004 alpaydn@boun.edu.r hp://www.cmpe.boun.edu.r/~ehem/2ml CHATER 3: Hdden Marov Models Inroducon Modelng dependences n npu; no

More information

Discrete Markov Process. Introduction. Example: Balls and Urns. Stochastic Automaton. INTRODUCTION TO Machine Learning 3rd Edition

Discrete Markov Process. Introduction. Example: Balls and Urns. Stochastic Automaton. INTRODUCTION TO Machine Learning 3rd Edition EHEM ALPAYDI he MI Press, 04 Lecure Sldes for IRODUCIO O Machne Learnng 3rd Edon alpaydn@boun.edu.r hp://www.cmpe.boun.edu.r/~ehem/ml3e Sldes from exboo resource page. Slghly eded and wh addonal examples

More information

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6)

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6) Econ7 Appled Economercs Topc 5: Specfcaon: Choosng Independen Varables (Sudenmund, Chaper 6 Specfcaon errors ha we wll deal wh: wrong ndependen varable; wrong funconal form. Ths lecure deals wh wrong ndependen

More information

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy Arcle Inernaonal Journal of Modern Mahemacal Scences, 4, (): - Inernaonal Journal of Modern Mahemacal Scences Journal homepage: www.modernscenfcpress.com/journals/jmms.aspx ISSN: 66-86X Florda, USA Approxmae

More information

ECE 366 Honors Section Fall 2009 Project Description

ECE 366 Honors Section Fall 2009 Project Description ECE 366 Honors Secon Fall 2009 Projec Descrpon Inroducon: Muscal genres are caegorcal labels creaed by humans o characerze dfferen ypes of musc. A muscal genre s characerzed by he common characerscs shared

More information

A Bayesian algorithm for tracking multiple moving objects in outdoor surveillance video

A Bayesian algorithm for tracking multiple moving objects in outdoor surveillance video A Bayesan algorhm for racng mulple movng obecs n oudoor survellance vdeo Manunah Narayana Unversy of Kansas Lawrence, Kansas manu@u.edu Absrac Relable racng of mulple movng obecs n vdes an neresng challenge,

More information

A Novel Iron Loss Reduction Technique for Distribution Transformers. Based on a Combined Genetic Algorithm - Neural Network Approach

A Novel Iron Loss Reduction Technique for Distribution Transformers. Based on a Combined Genetic Algorithm - Neural Network Approach A Novel Iron Loss Reducon Technque for Dsrbuon Transformers Based on a Combned Genec Algorhm - Neural Nework Approach Palvos S. Georglaks Nkolaos D. Doulams Anasasos D. Doulams Nkos D. Hazargyrou and Sefanos

More information

Li An-Ping. Beijing , P.R.China

Li An-Ping. Beijing , P.R.China A New Type of Cpher: DICING_csb L An-Png Bejng 100085, P.R.Chna apl0001@sna.com Absrac: In hs paper, we wll propose a new ype of cpher named DICING_csb, whch s derved from our prevous sream cpher DICING.

More information

( ) () we define the interaction representation by the unitary transformation () = ()

( ) () we define the interaction representation by the unitary transformation () = () Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger

More information

On computing differential transform of nonlinear non-autonomous functions and its applications

On computing differential transform of nonlinear non-autonomous functions and its applications On compung dfferenal ransform of nonlnear non-auonomous funcons and s applcaons Essam. R. El-Zahar, and Abdelhalm Ebad Deparmen of Mahemacs, Faculy of Scences and Humanes, Prnce Saam Bn Abdulazz Unversy,

More information

FTCS Solution to the Heat Equation

FTCS Solution to the Heat Equation FTCS Soluon o he Hea Equaon ME 448/548 Noes Gerald Reckenwald Porland Sae Unversy Deparmen of Mechancal Engneerng gerry@pdxedu ME 448/548: FTCS Soluon o he Hea Equaon Overvew Use he forward fne d erence

More information

Genetic Algorithm in Parameter Estimation of Nonlinear Dynamic Systems

Genetic Algorithm in Parameter Estimation of Nonlinear Dynamic Systems Genec Algorhm n Parameer Esmaon of Nonlnear Dynamc Sysems E. Paeraks manos@egnaa.ee.auh.gr V. Perds perds@vergna.eng.auh.gr Ah. ehagas kehagas@egnaa.ee.auh.gr hp://skron.conrol.ee.auh.gr/kehagas/ndex.hm

More information

The Analysis of the Thickness-predictive Model Based on the SVM Xiu-ming Zhao1,a,Yan Wang2,band Zhimin Bi3,c

The Analysis of the Thickness-predictive Model Based on the SVM Xiu-ming Zhao1,a,Yan Wang2,band Zhimin Bi3,c h Naonal Conference on Elecrcal, Elecroncs and Compuer Engneerng (NCEECE The Analyss of he Thcknesspredcve Model Based on he SVM Xumng Zhao,a,Yan Wang,band Zhmn B,c School of Conrol Scence and Engneerng,

More information

Detection of Waving Hands from Images Using Time Series of Intensity Values

Detection of Waving Hands from Images Using Time Series of Intensity Values Deecon of Wavng Hands from Images Usng Tme eres of Inensy Values Koa IRIE, Kazunor UMEDA Chuo Unversy, Tokyo, Japan re@sensor.mech.chuo-u.ac.jp, umeda@mech.chuo-u.ac.jp Absrac Ths paper proposes a mehod

More information

Object Tracking Based on Visual Attention Model and Particle Filter

Object Tracking Based on Visual Attention Model and Particle Filter Inernaonal Journal of Informaon Technology Vol. No. 9 25 Objec Trackng Based on Vsual Aenon Model and Parcle Fler Long-Fe Zhang, Yuan-Da Cao 2, Mng-Je Zhang 3, Y-Zhuo Wang 4 School of Compuer Scence and

More information

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair TECHNI Inernaonal Journal of Compung Scence Communcaon Technologes VOL.5 NO. July 22 (ISSN 974-3375 erformance nalyss for a Nework havng Sby edundan Un wh ang n epar Jendra Sngh 2 abns orwal 2 Deparmen

More information

Comparison of Differences between Power Means 1

Comparison of Differences between Power Means 1 In. Journal of Mah. Analyss, Vol. 7, 203, no., 5-55 Comparson of Dfferences beween Power Means Chang-An Tan, Guanghua Sh and Fe Zuo College of Mahemacs and Informaon Scence Henan Normal Unversy, 453007,

More information

A Deterministic Algorithm for Summarizing Asynchronous Streams over a Sliding Window

A Deterministic Algorithm for Summarizing Asynchronous Streams over a Sliding Window A Deermnsc Algorhm for Summarzng Asynchronous Sreams over a Sldng ndow Cosas Busch Rensselaer Polyechnc Insue Srkana Trhapura Iowa Sae Unversy Oulne of Talk Inroducon Algorhm Analyss Tme C Daa sream: 3

More information

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue. Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons

More information

Anomaly Detection. Lecture Notes for Chapter 9. Introduction to Data Mining, 2 nd Edition by Tan, Steinbach, Karpatne, Kumar

Anomaly Detection. Lecture Notes for Chapter 9. Introduction to Data Mining, 2 nd Edition by Tan, Steinbach, Karpatne, Kumar Anomaly eecon Lecure Noes for Chaper 9 Inroducon o aa Mnng, 2 nd Edon by Tan, Senbach, Karpane, Kumar 2/14/18 Inroducon o aa Mnng, 2nd Edon 1 Anomaly/Ouler eecon Wha are anomales/oulers? The se of daa

More information

Introduction to Boosting

Introduction to Boosting Inroducon o Boosng Cynha Rudn PACM, Prnceon Unversy Advsors Ingrd Daubeches and Rober Schapre Say you have a daabase of news arcles, +, +, -, -, +, +, -, -, +, +, -, -, +, +, -, + where arcles are labeled

More information

2. SPATIALLY LAGGED DEPENDENT VARIABLES

2. SPATIALLY LAGGED DEPENDENT VARIABLES 2. SPATIALLY LAGGED DEPENDENT VARIABLES In hs chaper, we descrbe a sascal model ha ncorporaes spaal dependence explcly by addng a spaally lagged dependen varable y on he rgh-hand sde of he regresson equaon.

More information

Off line signatures Recognition System using Discrete Cosine Transform and VQ/HMM

Off line signatures Recognition System using Discrete Cosine Transform and VQ/HMM Ausralan Journal of Basc and Appled Scences, 6(12): 423-428, 2012 ISSN 1991-8178 Off lne sgnaures Recognon Sysem usng Dscree Cosne Transform and VQ/HMM 1 Behrouz Vasegh, 2 Somayeh Hashem, 1,2 Deparmen

More information

CHAPTER 10: LINEAR DISCRIMINATION

CHAPTER 10: LINEAR DISCRIMINATION CHAPER : LINEAR DISCRIMINAION Dscrmnan-based Classfcaon 3 In classfcaon h K classes (C,C,, C k ) We defned dscrmnan funcon g j (), j=,,,k hen gven an es eample, e chose (predced) s class label as C f g

More information

EP2200 Queuing theory and teletraffic systems. 3rd lecture Markov chains Birth-death process - Poisson process. Viktoria Fodor KTH EES

EP2200 Queuing theory and teletraffic systems. 3rd lecture Markov chains Birth-death process - Poisson process. Viktoria Fodor KTH EES EP Queung heory and eleraffc sysems 3rd lecure Marov chans Brh-deah rocess - Posson rocess Vora Fodor KTH EES Oulne for oday Marov rocesses Connuous-me Marov-chans Grah and marx reresenaon Transen and

More information

Stochastic Repair and Replacement with a single repair channel

Stochastic Repair and Replacement with a single repair channel Sochasc Repar and Replacemen wh a sngle repar channel MOHAMMED A. HAJEEH Techno-Economcs Dvson Kuwa Insue for Scenfc Research P.O. Box 4885; Safa-309, KUWAIT mhajeeh@s.edu.w hp://www.sr.edu.w Absrac: Sysems

More information

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC CH.3. COMPATIBILITY EQUATIONS Connuum Mechancs Course (MMC) - ETSECCPB - UPC Overvew Compably Condons Compably Equaons of a Poenal Vecor Feld Compably Condons for Infnesmal Srans Inegraon of he Infnesmal

More information

Sampling Procedure of the Sum of two Binary Markov Process Realizations

Sampling Procedure of the Sum of two Binary Markov Process Realizations Samplng Procedure of he Sum of wo Bnary Markov Process Realzaons YURY GORITSKIY Dep. of Mahemacal Modelng of Moscow Power Insue (Techncal Unversy), Moscow, RUSSIA, E-mal: gorsky@yandex.ru VLADIMIR KAZAKOV

More information

Bayes rule for a classification problem INF Discriminant functions for the normal density. Euclidean distance. Mahalanobis distance

Bayes rule for a classification problem INF Discriminant functions for the normal density. Euclidean distance. Mahalanobis distance INF 43 3.. Repeon Anne Solberg (anne@f.uo.no Bayes rule for a classfcaon problem Suppose we have J, =,...J classes. s he class label for a pxel, and x s he observed feaure vecor. We can use Bayes rule

More information

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading Onlne Supplemen for Dynamc Mul-Technology Producon-Invenory Problem wh Emssons Tradng by We Zhang Zhongsheng Hua Yu Xa and Baofeng Huo Proof of Lemma For any ( qr ) Θ s easy o verfy ha he lnear programmng

More information

Video-Based Face Recognition Using Adaptive Hidden Markov Models

Video-Based Face Recognition Using Adaptive Hidden Markov Models Vdeo-Based Face Recognon Usng Adapve Hdden Markov Models Xaomng Lu and suhan Chen Elecrcal and Compuer Engneerng, Carnege Mellon Unversy, Psburgh, PA, 523, U.S.A. xaomng@andrew.cmu.edu suhan@cmu.edu Absrac

More information

Effect of Resampling Steepness on Particle Filtering Performance in Visual Tracking

Effect of Resampling Steepness on Particle Filtering Performance in Visual Tracking 102 The Inernaonal Arab Journal of Informaon Technology, Vol. 10, No. 1, January 2013 Effec of Resamplng Seepness on Parcle Flerng Performance n Vsual Trackng Zahdul Islam, Ch-Mn Oh, and Chl-Woo Lee School

More information

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study)

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study) Inernaonal Mahemacal Forum, Vol. 8, 3, no., 7 - HIKARI Ld, www.m-hkar.com hp://dx.do.org/.988/mf.3.3488 New M-Esmaor Objecve Funcon n Smulaneous Equaons Model (A Comparave Sudy) Ahmed H. Youssef Professor

More information

Epistemic Game Theory: Online Appendix

Epistemic Game Theory: Online Appendix Epsemc Game Theory: Onlne Appendx Edde Dekel Lucano Pomao Marcano Snscalch July 18, 2014 Prelmnares Fx a fne ype srucure T I, S, T, β I and a probably µ S T. Le T µ I, S, T µ, βµ I be a ype srucure ha

More information

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue. Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm H ( q, p, ) = q p L( q, q, ) H p = q H q = p H = L Equvalen o Lagrangan formalsm Smpler, bu

More information

( ) [ ] MAP Decision Rule

( ) [ ] MAP Decision Rule Announcemens Bayes Decson Theory wh Normal Dsrbuons HW0 due oday HW o be assgned soon Proec descrpon posed Bomercs CSE 90 Lecure 4 CSE90, Sprng 04 CSE90, Sprng 04 Key Probables 4 ω class label X feaure

More information

A Novel Object Detection Method Using Gaussian Mixture Codebook Model of RGB-D Information

A Novel Object Detection Method Using Gaussian Mixture Codebook Model of RGB-D Information A Novel Objec Deecon Mehod Usng Gaussan Mxure Codebook Model of RGB-D Informaon Lujang LIU 1, Gaopeng ZHAO *,1, Yumng BO 1 1 School of Auomaon, Nanjng Unversy of Scence and Technology, Nanjng, Jangsu 10094,

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm Hqp (,,) = qp Lqq (,,) H p = q H q = p H L = Equvalen o Lagrangan formalsm Smpler, bu wce as

More information

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data Anne Chao Ncholas J Goell C seh lzabeh L ander K Ma Rober K Colwell and Aaron M llson 03 Rarefacon and erapolaon wh ll numbers: a framewor for samplng and esmaon n speces dversy sudes cology Monographs

More information

Lecture 11 SVM cont

Lecture 11 SVM cont Lecure SVM con. 0 008 Wha we have done so far We have esalshed ha we wan o fnd a lnear decson oundary whose margn s he larges We know how o measure he margn of a lnear decson oundary Tha s: he mnmum geomerc

More information

Constrained-Storage Variable-Branch Neural Tree for. Classification

Constrained-Storage Variable-Branch Neural Tree for. Classification Consraned-Sorage Varable-Branch Neural Tree for Classfcaon Shueng-Ben Yang Deparmen of Dgal Conen of Applcaon and Managemen Wenzao Ursulne Unversy of Languages 900 Mnsu s oad Kaohsng 807, Tawan. Tel :

More information

Department of Economics University of Toronto

Department of Economics University of Toronto Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of

More information

OMXS30 Balance 20% Index Rules

OMXS30 Balance 20% Index Rules OMX30 Balance 0% ndex Rules Verson as of 30 March 009 Copyrgh 008, The NADAQ OMX Group, nc. All rghs reserved. NADAQ OMX, The NADAQ ock Marke and NADAQ are regsered servce/rademarks of The NADAQ OMX Group,

More information

II. Light is a Ray (Geometrical Optics)

II. Light is a Ray (Geometrical Optics) II Lgh s a Ray (Geomercal Opcs) IIB Reflecon and Refracon Hero s Prncple of Leas Dsance Law of Reflecon Hero of Aleandra, who lved n he 2 nd cenury BC, posulaed he followng prncple: Prncple of Leas Dsance:

More information

CHAPTER 7: CLUSTERING

CHAPTER 7: CLUSTERING CHAPTER 7: CLUSTERING Semparamerc Densy Esmaon 3 Paramerc: Assume a snge mode for p ( C ) (Chapers 4 and 5) Semparamerc: p ( C ) s a mure of denses Mupe possbe epanaons/prooypes: Dfferen handwrng syes,

More information

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment EEL 6266 Power Sysem Operaon and Conrol Chaper 5 Un Commmen Dynamc programmng chef advanage over enumeraon schemes s he reducon n he dmensonaly of he problem n a src prory order scheme, here are only N

More information

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s Ordnary Dfferenal Equaons n Neuroscence wh Malab eamples. Am - Gan undersandng of how o se up and solve ODE s Am Undersand how o se up an solve a smple eample of he Hebb rule n D Our goal a end of class

More information

doi: info:doi/ /

doi: info:doi/ / do: nfo:do/0.063/.322393 nernaonal Conference on Power Conrol and Opmzaon, Bal, ndonesa, -3, June 2009 A COLOR FEATURES-BASED METHOD FOR OBJECT TRACKNG EMPLOYNG A PARTCLE FLTER ALGORTHM Bud Sugand, Hyoungseop

More information

Local Cost Estimation for Global Query Optimization in a Multidatabase System. Outline

Local Cost Estimation for Global Query Optimization in a Multidatabase System. Outline Local os Esmaon for Global uery Opmzaon n a Muldaabase ysem Dr. ang Zhu The Unversy of Mchgan - Dearborn Inroducon Oulne hallenges for O n MDB uery amplng Mehod ualave Approach Fraconal Analyss and Probablsc

More information

Algorithm Research on Moving Object Detection of Surveillance Video Sequence *

Algorithm Research on Moving Object Detection of Surveillance Video Sequence * Opcs and Phooncs Journal 03 3 308-3 do:0.436/opj.03.3b07 Publshed Onlne June 03 (hp://www.scrp.org/journal/opj) Algorhm Research on Movng Objec Deecon of Survellance Vdeo Sequence * Kuhe Yang Zhmng Ca

More information

CRAWLING DEEP WEB CONTENT THROUGH QUERY FORMS

CRAWLING DEEP WEB CONTENT THROUGH QUERY FORMS CRAWLING DEEP WEB CONTENT THROUGH QUERY FORMS Jun Lu, Zhaohu Wu, Lu Jang, Qnghua Zheng, Xao Lu MOE KLINNS Lab and SKLMS Lab, X an Jaoong Unversy, X'an 70049, Chna lueen@mal.xu.edu.cn, wzh@su.xu.edu.cn,

More information

ELASTIC MODULUS ESTIMATION OF CHOPPED CARBON FIBER TAPE REINFORCED THERMOPLASTICS USING THE MONTE CARLO SIMULATION

ELASTIC MODULUS ESTIMATION OF CHOPPED CARBON FIBER TAPE REINFORCED THERMOPLASTICS USING THE MONTE CARLO SIMULATION THE 19 TH INTERNATIONAL ONFERENE ON OMPOSITE MATERIALS ELASTI MODULUS ESTIMATION OF HOPPED ARBON FIBER TAPE REINFORED THERMOPLASTIS USING THE MONTE ARLO SIMULATION Y. Sao 1*, J. Takahash 1, T. Masuo 1,

More information

(,,, ) (,,, ). In addition, there are three other consumers, -2, -1, and 0. Consumer -2 has the utility function

(,,, ) (,,, ). In addition, there are three other consumers, -2, -1, and 0. Consumer -2 has the utility function MACROECONOMIC THEORY T J KEHOE ECON 87 SPRING 5 PROBLEM SET # Conder an overlappng generaon economy le ha n queon 5 on problem e n whch conumer lve for perod The uly funcon of he conumer born n perod,

More information